Piloting a food photo sorting activity in Samoa to assess maternal beliefs and their role in child diet

Abstract

1. INTRODUCTION

Samoa, a small island nation in the South Pacific, faces an alarming and rising burden of obesity, metabolic syndrome, and associated noncommunicable diseases (NCDs; Hawley & McGarvey, 2015; Ng et al., 2014; World Health Organization, 2002). A survey of 25‐ to 64‐year‐old Samoan adults over the past three decades revealed that obesity (body mass index [BMI] ≥30 kg/m²) rose from 27.7% to 53.1% among men and 44.4% to 76.7% among women, while the prevalence of Type II diabetes rose from 1.2% to 19.6% among men and 2.2% to 19.5% among women (Lin et al., 2017). Although the increasing prevalence has been well‐documented among adults, the burden of overweight/obesity has been less studied among Samoan children. Recent data from the Ola Tuputupua'e (Growing Up) study, following the growth and development of a cohort of ~500 Samoan children, estimated that 16–30% of 2‐ to 4‐year‐olds were overweight/obese (Choy et al., 2017), indicating that interventions to reduce obesity risk must begin early in life.

Researchers have attributed the increase in prevalence of obesity and related NCDs among the Samoan population to a nutritional transition, characterized by urbanization, a shift towards sedentary lifestyle and employment, and modernization of traditional diets (Keighley et al., 2007; Popkin, 1998). In the past half century, food imports to Samoa have increased fivefold resulting in an influx of processed foods (high in energy density, refined sugar, white flour, trans fat, polyunsaturated fats, and salt) and a 47% increase in available energy from food (DiBello et al., 2009; FAOSTAT, 2002; Seiden, Hawley, Schulz, Raifman, & McGarvey, 2012; Shridhar et al., 2015). Following a ‘modern’ diet has been associated with obesity and metabolic syndrome in Samoan adults and obesity in children (Choy, 2018; DiBello et al., 2009; Wang et al., 2017).

A systematic literature review of 25 studies around the world concluded that children who are overweight/obese have an increased risk of remaining overweight/obese as adults (Singh, Mulder, Twisk, Van Mechelen, & Chinapaw, 2008). This trend may reflect the tendency for dietary habits formed in early childhood to persist into late childhood and adolescence (Kelder, Perry, Klepp, & Lytle, 1994; Skinner, Carruth, Bounds, & Ziegler, 2002). Early childhood is also a unique transitional period during which parents hold great influence over their child's food behaviours and diet (e.g., food selection, preferences, and portion sizes; Scaglioni, Salvioni, & Galimberti, 2008; Watkins & Jones, 2015). Understanding the factors that influence parental feeding behaviour is essential to shaping a healthy early child diet. A study in neighbouring American Samoa explored mothers' attitudes and beliefs about infant feeding, with a particular focus on breastfeeding (Hawley et al., 2015), but mothers' attitudes and beliefs as they pertain to early childhood diet have yet to be explored.

Studies conducted in other settings have focused on maternal nutritional knowledge, finding some associations with child diet, but few have examined the role of additional factors such as the cost and social value of foods. Among Omani and Flemish mothers, lower maternal nutritional knowledge was associated with lower dietary adequacy for preschool‐aged children (Al‐Shookri, Al‐Shukaily, Hassan, Al‐Sheraji, & Al‐Tobi, 2011; Vereecken & Maes, 2010). In Mozambique, high maternal nutritional knowledge led to a more diversified diet for their preschool‐aged child (Burchi, 2010). In Ghana and Lesotho, maternal nutritional knowledge of infant feeding was associated with infant nutritional status (Appoh & Krekling, 2005; Ruel, Habicht, Pinstrup‐Andersen, & Gröhn, 1992). However, in socio‐economically disadvantaged neighbourhoods in Victoria, Australia, researchers were unable to find an association between maternal nutritional knowledge and preschool‐aged child diet; they hypothesized that limited resources and cost of foods may have played a larger role (Williams, Campbell, Abbott, Crawford, & Ball, 2012). Understanding child diet is additionally complex in Samoa because meals can be shared with nuclear and extended family members sometimes in multiple households. In this context, additional structural and cultural factors may influence what Samoan mothers feed their children.

To address current gaps in knowledge and inform future intervention strategies, the following study examines maternal beliefs surrounding food healthfulness, social status, and cost as well as maternal nutritional and consumer knowledge as they relate to early child dietary intake.

2. METHODS

3. RESULTS

3.2. Maternal beliefs surrounding the healthfulness, social status, and cost of traditional foods

Aggregating all mothers' rankings revealed that, as a group, the mothers believed fruits and vegetables were the healthiest food item, while viewing soda as the least healthy (Table ​(Table2).2). Across the entire sample of mothers, boiled fish was ranked as the healthiest source of protein whereas fried chicken was the least healthy and taro, closely followed by breadfruit, as the healthiest source of complex carbohydrates whereas fries was the least healthy. The foods considered to have the highest social status were oranges, followed by papaya and apple, whereas the lowest were ice cream, fried pancake balls, and soda. The foods perceived to be the most expensive were pizza, fried chicken, and fries, whereas the least expensive were banana, papaya, laupele, and water. On the island, taro, breadfruit, carrots, eggplant, pumpkin, laupele, bananas, and papaya are grown commonly on home plantations and often consumed at no monetary cost for families who have home plantations (Samoa Bureau of Statistics, 2018b). This was reflected in the aggregate cost rankings. Traditional foods, as defined by DiBello et al. (2009) are marked with letters in Table ​Table22.

Table 2

Mother's (n = 44) average rankings of frequently consumed foods^a

	Healthfulness	Social status	Perceived cost
Highest	Fruits and vegetablesa	Orange	Pizza
	Boiled fisha	Papayaa	Fried chicken
	Taroa	Apple	Fries
	Breadfruita	Bananaa	Grilled chicken
	Water	Taroa	Fried fish
	White rice	Boiled fisha	Boiled fisha
	Fried fish	Pumpkina	Tinned fish
	Eggs	Carrotsa	Ice cream
	Tinned fish	Peppers	White rice
	White bread	Laupelea	Carrotsa
	Tinned fish in tomato sauce	Eggplanta	Eggs
	Grilled chicken	Breadfruita	Soda
	Tea and coffee	Fried fish	Tinned fish in tomato sauce
	Pizza	Fried chicken	Peppers
	Fried pancake balls	Grilled Chicken	Pumpkina
	Fried chicken	Pizza	Orange
	Fries	Water	Apple
	Ice cream	Eggs	Taroa
	Soda	White Rice	Eggplanta
	—	Fries	Fried pancake balls
	—	Tinned fish	White bread
	—	Tinned fish in tomato sauce	Breadfruita
	—	White bread	Tea and coffee
	—	Tea and coffee	Bananaa
	—	Ice cream	Papayaa
	—	Fried pancake balls	Laupelea
Lowest	—	Soda	Water

Open in a separate window

^aTraditional foods.

Bivariate analysis of the aggregate rankings revealed that mothers' ranking of food healthfulness was strongly, positively correlated with social status (r = .87, 95% CI [0.72, 0.94], p < .001) and moderately, negatively correlated with perceived cost (r = −.45, 95% CI [−0.71, −0.09], p = .02; Figure ​Figure1a,b).1a,b). Social status and perceived costs were not statistically significantly correlated (r = −.17, 95% CI [−0.51, 0.23], p = .41; Figure ​Figure1c).1c). Locating traditional foods in these correlations reveal that traditional foods are viewed as healthier, of higher social status, and of lower perceived cost than non‐traditional/imported ones.

Open in a separate window

Figure 1

The average of mother's (n = 44) rankings of each food item in each sorting scheme are plotted. The trend line represents the correlation between average mother's (a) health and social status ranking (r = .87, 95% CI [0.72, 0.94], p < .001), (b) health and cost rankings (r = −.45, 95% CI [−0.71, −0.09], p = .02), and (c) social status and cost rankings (r = −.17, 95% CI [−0.51, 0.23], p = .41)

3.3. Maternal nutritional and consumer price knowledge

Mothers' average health rankings (r =0.87, 95% CI [0.68, 0.95], p < .001) and cost rankings (r = .84, 95% CI [0.68, 0.93], p < .001) were strongly, positively correlated with the rankings made by nutritional experts and the market survey, respectively (Figure ​(Figure2).2). However, mothers tended to rank bread, fried pancake, pizza, and taro as healthier and eggs, grilled chicken, and water as less healthy than experts did (Figure S1a). They also tended to overestimate the cost of bananas, carrots, papayas, rice, and taro, while underestimating the cost of apples, coffee/tea, eggs, fries, oranges, pizza, and tinned fish (Figure S1b).

Open in a separate window

Figure 2

(a) Correlation between average mother's (n = 44) and average expert's (n = 8) health rankings (r = .87, 95% CI [0.68, 0.95], p < .001). (b) Correlation between average mother's (n = 44) food cost rankings and a cost by volume (grams) ranking determined by market survey (r = .84, 95% CI [0.68, 0.93], p < .001)

3.4. Correlation between mother's food beliefs and their child's diet

Mothers' perceived food cost (r = −.37, 95% CI [−0.66, 0.02], p = .06) was related to their child's usual frequency of food intake (Figure ​(Figure3a),3a), but food healthfulness (r = 0.10, 95% CI [−0.40, 0.38], p = .95) or social status (r = −0.25, 95% CI [−0.58, 0.15], p = .21) were not (Figure ​(Figure3b,c).3b,c). The more expensive mothers ranked a food item, the less frequently they reported that their child ate it. Foods that did not follow this pattern were bread, tea, and white rice. Bivariate correlations of individual mother's rankings of each food item and percentage of their child's daily caloric intake from each food item revealed limited significant associations (Table S1).

Open in a separate window

Figure 3

(a) Correlation between average mother's (n = 44) health rankings (r = −.10, 95% CI [−0.40, 0.38], p = .95), (b) social status rankings (r = −.25, 95% CI [−0.58, 0.15], p = .21), and (c) cost rankings (r = −.37, 95% CI [−0.66, 0.02], p = .06) and average child's (n = 44) weekly frequency of food intake

4. DISCUSSION

Despite the fact that child food preferences are often established early in life and that parental beliefs and attitudes towards food likely play a major role in establishing those preferences, there had been no previous attempt to characterize parental beliefs about food in Samoa, a setting with prevalent childhood obesity. This study addresses that gap by documenting beliefs about commonly consumed foods and nutritional knowledge among Samoan mothers. Repurposing a previously qualitative food sorting task, we were able to demonstrate how maternal beliefs were related to child dietary intake, noting that (a) traditional foods were thought to be healthier, of higher social status, and lower cost than non‐traditional/imported foods, (b) that mothers had accurate nutritional and consumer knowledge, and (c) that cost of food influenced child intake, whereas healthfulness and social status did not.

The increasing consumption of a modern diet, enabled by the fivefold increase in imported foods over the past half century, suggested that imported foods had attained a higher social status than traditional foods among Samoans (DiBello et al., 2009; FAOSTAT, 2002). The mothers' contradictory ranking of traditional foods as healthier and of higher social status may be a product of recent eating health and economic promotion campaigns in both Samoa and the wider Pacific region where there has been a renewed focus on consumption of local produce for health. The Samoan Ministry of Health's 2016 “Eat the Rainbow” initiative encourages the consumption of a variety of local fruits and vegetables, and the Samoan Women in Business Development endorses organic produce to stimulate the local economy (“Samoa Observer, 2016”; “Women in Business Samoa, n.d.”; World Bank, 2018). In addition, the Secretariat for the Pacific Community recently released the first Pacific Guidelines for a Healthy Diet and Lifestyle, which placed emphasis on consuming local produce and traditionally grown foods for good health (Public Health Division of the Pacific Community, 2017).

This study finds that, on average, Samoan mothers have largely accurate knowledge of the relative healthfulness and cost of staple food items for the most part independent of mothers' highest level of educational attainment. Therefore, lack of knowledge seems to not be a barrier to healthy child diet. However, examining individual mother's rankings revealed some potential gaps in nutritional and consumer knowledge, which could still inform future public health action.

Mothers tended to overestimate the healthfulness of carbohydrate‐rich foods, in particular, bread and pizza compared with expert rankings. Mothers also tended to overestimate the healthfulness of fried foods (fried pancake balls) while underestimating the healthfulness of unfried foods (grilled chicken). However, alternative explanations may exist. For instance, grilling is an uncommon food preparation method and therefore, mothers may be unfamiliar with it. In general, the stock photos used for the sorting activity may have looked unfamiliar; however, the photos were selected with the help of our local enumerators. Another alternative explanation could also be portion size. While the healthfulness of fried pancake balls was overestimated, the healthfulness of fried chicken was more accurate. A local nutritional expert explained that Samoans view fried pancake balls as a light snack and therefore may be misestimating the amount of oil they consume when eating it. Public health education on the healthfulness of processed versus unprocessed carbohydrates and methods of food preparation may be beneficial.

In terms of cost misconception, mothers tended to overestimate the price of local produce while underestimating the price of imported produce. This suggests that efforts could be made to inform mothers of the cost effectiveness of growing and/or buying locally grown foods where time and other resources allow. A pilot study completed among mothers and children living in socio‐economically disadvantaged neighbourhoods in Victoria, Australia, demonstrated that a brief 10‐min slideshow was an effective educational intervention targeting perceptions of healthy food affordability (Williams et al., 2014). This intervention could be adapted for use in Samoa. However, one limitation to this specific analysis that should be noted is that we observed mothers collectively placing fruit and vegetable cards at the lowest end of the cost spectrum. They sometimes sorted between them arbitrarily or ranked them as equal in cost. Yet another alternative explanation is the price instability of local produce due to seasonal variations and periodic natural disasters (Samoan Ministry of Agriculture and Fisheries, 2011, 2016; Seiden et al., 2012).

In this sample, cost, but not healthfulness or social status, was modestly correlated with child diet. In this setting where the average weekly expenditure on food is $65.35 WST per capita per week compared with the average income that ranges from $25.60 to $103.70 WST per capita per week, costs of food, and not healthfulness or social status, may be the most salient factor in mothers' feeding decisions. This inverse relationship between food cost and frequency of child intake aligns with the existing literature, particularly of low socio‐economic status populations. Previous qualitative studies have identified perceived food cost as a primary factor influencing healthy eating and food choice among low‐income households (Darmon & Drewnowski, 2015; Inglis, Ball, & Crawford, 2005), and an inverse relationship between food prices and intake has been found in studies across North America, Europe, and Australia (Lee, Ralston, & Truby, 2011). In fact, other studies suggest that food cost mediates the association between low socio‐economic status and poor diet quality (Aggarwal, Monsivais, Cook, & Drewnowski, 2011; Inglis et al., 2005). Future studies might examine if this phenomenon is occurring in Samoa.

The true strength of the correlation between perceived food cost and child diet may be underestimated by the presence of outlying foods—white rice, bread, and tea—which seem to be most highly consumed among children regardless of their perceived cost. The cost misconception plots reveal that the cost of white rice—the greatest “outlier”—is overestimated by approximately 86% of Samoan mothers sampled (Figure S1b). This may be because white rice is sold in bulk, causing mothers to recall the price of an entire bag rather than a single serving. If white rice were removed from analysis, the correlation between average maternal cost ranking and child intake would become much stronger (r = −.53, 95% CI [−0.76, −0.16], p = .007).

The negative correlation between food cost and child intake seems to be discordant with the finding that traditional foods were ranked as healthier and less expensive, suggesting that cost was not a barrier to accessing healthy foods. Many of the least expensive foods were those which are commonly grown at little to no cost in home plantations. One explanation for this discrepancy may be that cost is only a barrier to accessing healthy foods, which are purchased rather than grown.

The paucity of correlations between individual mother's rankings of food healthfulness, social status, and cost and their child's intake of the food item might mean that other unmeasured variables were at play. In other studies, household availability of fruits and vegetables, salty foods, and soft drinks have been found to influence children's intake of these foods (Campbell et al., 2013; Cullen et al., 2001). Parental rules, modelling, and restrictions have also been shown to influence child dietary intake; rules around eating a certain amount of fruits and vegetables and maternal consumption of fruits and vegetables have a positive effect on child consumption while a systematic literature review found that parental feeding restrictions are associated with increased eating and childhood weight (Faith, Scanlon, Birch, Francis, & Sherry, 2004; Jones, Steer, Rogers, & Emmett, 2010).

The novel quantitative analysis of this food photo sorting activity has great potential for use in other populations. It could assess food beliefs in other contexts by simply replacing the staple Samoan food items with those that are most common in the population of interest. Additional culturally relevant maternal belief systems, beyond the ones measured here, could also be assessed by changing the sorting prompts. This tool, first developed by Hesketh et al. (2005) for use among children, is completely pictorial with no literacy requirements therefore making it accessible to almost all participants. The straightforward analysis of food rankings has the potential to further the understanding of maternal food beliefs and nutritional knowledge in populations for which little is known and who may bear the highest burdens of overweight/obesity and NCDs.

5. CONCLUSION

Using a simple food photo sorting activity, we were able to document maternal beliefs about the nutritional value, social status, and cost of foods commonly consumed by Samoan children. Although we provide evidence that locally grown, traditional Samoan foods are favoured by mothers and believed to be healthier, only the cost of food was associated with child dietary intake, indicating the continued promotion as a strategy for improving child nutritional intake must be supported with actions to increase access to reasonably priced healthy food options.

CONFLICTS OF INTEREST

The authors declare that they have no conflicts of interest.

CONTRIBUTIONS

VKT was responsible for study design under the supervision of NLH. VKT, CCC, and TJA conducted the research with the support of TN, CS and MSR. VKT, and NLH analysed and interpreted the data. VKT drafted the manuscript. All authors reviewed and edited the final manuscript.

Supporting information

ACKNOWLEDGMENTS

Notes

Tanawattanacharoen VK, Choy CC, Anesi TJ, et al. Piloting a food photo sorting activity in Samoa to assess maternal beliefs and their role in child diet. Matern Child Nutr. 2020;16:e12974 10.1111/mcn.12974 [PMC free article] [PubMed] [CrossRef] [Google Scholar]

Contributor Information

Veeraya K. Tanawattanacharoen, Email: ude.elay@neorahcanattawanat.ayareev.

Nicola L. Hawley, Email: ude.elay@yelwah.alocin.

REFERENCES