Biocultural Determinants of Mothers' Complementary Feeding Decisions in the Urban Brazilian Amazon

ABSTRACT

Objective

Complementary feeding (CF) occurs during a critical period of infant growth and development with life‐long health implications. Despite international efforts, there remains significant variation in the adequacy of complementary diets across settings. In Brazil, there is marked variation in adherence to CF guidelines and infant growth outcomes, with the north (Amazon) lagging other regions of the country. This study aimed to characterize the complementary diet and develop a model to explain feeding decisions in the Amazonian city of Belém.

Methods

With a sample of n = 30 mothers, we combined 24‐h dietary recalls, a pile‐sorting activity, and in‐depth interviews to address study aims. Using descriptive statistics, we analyzed the pile‐sort data to characterize the evolving complementary diet. Then, using thematic analysis of interview transcripts, we identified the most salient factors shaping mothers' feeding decisions.

Results

While there was variation in opinion regarding the timing of introduction of liquids other than breastmilk and ultra‐processed convenience foods, we found high consensus regarding the ideal complementary diet which, beginning at 6 months, met WHO dietary diversity guidelines and evolved with infant age. Three themes—integrating and applying trusted sources of advice, infant readiness and future health, and challenges to feeding ideals—illustrate how socioeconomic, cultural, and infant bio‐behavioral cues interact to shape CF.

Conclusion

Efforts to improve infant feeding must move beyond identifying individual factors and toward biocultural models that consider how political–economic and local contexts interact to influence the ethnomedical systems, household sociocultural dynamics, including income, gender, and age‐based responsibilities, and power relations that shape feeding behaviors.

1. Introduction

Complementary feeding is defined as the period when foods and beverages are introduced into infants' diets to supplement breastfeeding. As a practice it includes the timing of introduction, the selection of foods and their preparation, quantity given, and feeding style and frequency (Stewart et al. 2013). The World Health Organization (WHO) (2023) recommends that complementary feeding begin when an infant reaches 6 months of age, a developmental stage when breastmilk can no longer meet the infant's nutritional needs, and that it continues until 23 months of age or the time when the infant is fully weaned and consuming the family diet.

The period between six and 23 months is a critical time as infants continue to experience rapid growth and development (Bogin 2020), meaning they need a varied, nutrient‐rich diet but are developmentally limited in what and how much they can consume (Dewey 2013; Dewey et al. 2021; Sellen 2007). A robust body of scholarship demonstrates that inadequate complementary feeding places infants at risk of growth faltering and poor health in both the short and long term (Black et al. 2008; Michaelsen and Friis 1998; Victora et al. 2010). For these reasons, international and national health agencies, using current knowledge of infant nutrition, develop and disseminate recommendations for complementary feeding (World Health Organization [WHO] 2023) and governments, through national policies, promote guidelines and monitor outcomes.

Current WHO recommendations focus on dietary diversity and meal frequency. For example, the WHO (2023) recommends that infants 6–23 months consume foods from different food groups, for example, animal source foods, fruits, vegetables, cereals, pulses, nuts, and seeds as a diverse diet is more likely to provide the macro‐ and micronutrients infants require to meet their growth and development needs (Choudhury et al. 2019; Dewey et al. 2021; Onyango et al. 2014). The WHO (2023) also recommends responsive feeding, that is, that the caregiver responds to infant cues indicating hunger and satiety as this approach encourages infant self‐regulation to support growth and development (see also Pelto 2000). Measures such as minimum dietary diversity (MDD) and minimal meal frequency (MMF) are used to monitor adherence to these recommendations (WHO 2023).

While the literature on complementary feeding is vast, it is dominated by national, regional, and interregional studies largely concerned with documenting adherence to WHO recommendations (e.g., Cavalcanti and Boccolini 2022; Dhami et al. 2019; Gatica‐Domínguez et al. 2021; Herman et al. 2023; White et al. 2017). These studies report significant inter‐ and intrapopulation variation in MDD and MMF and persistent challenges meeting international recommendations. While few attempt to explain observed variation, those that do focus on a limited number of household and individual‐level variables included in the surveys upon which they are based and find income and maternal education positively associated with MDD, presumably via food access and choice.

Numerous scholars have argued that to understand variation in and design effective policies to improve infants' diets, complementary feeding must be appreciated as not only a biological necessity but as part of constantly evolving cultural systems that give it meaning and regulate its practice (Engle et al. 1999; Jelliffe 1962; Macadam and Dettwyler 1995; Mwaseba et al. 2016; Pelto et al. 2003; Pelto and Pelto 1997). Others add that the influence of the broader political‐economic context in which local systems are embedded must also be considered (Baker et al. 2023; Hawkes et al. 2020; Popkin 1986; Popkin and Solon 1976). In response, researchers have conducted ethnographically grounded studies aimed at gaining an emic perspective on infant feeding and, in doing so, have expanded our understanding of the factors influencing complementary feeding decisions. While many identify household income as important (e.g., Dutta et al. 2006; Lindsay et al. 2009; Matvienko‐Sikar et al. 2018; Monterrosa et al. 2012; Nogueira et al. 2022; Rinaldi and Conde 2019), they also suggest other factors play a role. Some highlight the importance of local food cultures, especially food classification systems such as the hot/cold system found across Latin America (Foster 1994; Messer 1981), as they inform people's perceptions of foods and, thus, their appropriateness for infants (e.g., Dutta et al. 2006; Monterossa et al. 2012; Rodriguez‐Oliveros et al. 2014; Zobrist et al. 2017). Others emphasize the role of caregivers' perceptions of infant health and development (i.e., infants' readiness) (Lindsay et al. 2009; Cheney et al. 2019; Rodriguez‐Oliveros et al. 2014) or behaviors interpreted as signs of hunger and/or interest (Dutta et al. 2006; Monterrosa et al. 2012) as driving feeding decisions. Finally, while large‐scale studies draw attention to the relationship between maternal education and infant dietary diversity, qualitative studies show that other individuals, particularly, female kin, can be influential in feeding decisions (e.g., Kerr et al. 2008; Cheney et al. 2019; Lindsay et al. 2009; Monterrosa et al. 2012; Rodriguez‐Oliveros et al. 2014). Overall, ethnographic studies support the idea that research on complementary feeding requires a biocultural approach, that is, recognition that infants' have an evolved developmental trajectory and related nutritional needs, which are attended to and, thus, influenced by the complex cultural and socioeconomic milieu in which feeding occurs (Dettwyler 2004; Kodish et al. 2019; Meehan and Roulette 2013; Moffat 2001; Palmer 2011; Sellen 2009).

Responding to calls for a more holistic approach, the goal of this biocultural study is to advance our understanding of complementary feeding and the factors that influence it in the Amazonian city of Belém, Pará, Brazil. Brazil is a nation with strong policies aimed at encouraging breastfeeding and the adoption of international complementary feeding recommendations (Brasil Ministério da Saúde Secretaria de Atenção à Saúde Departamento de Ações Programáticas Estratégicas 2017) but also a place of ecological and cultural diversity, marked inequality, and experiencing a surge in access to and utilization of ultra‐processed foods (Baker et al. 2021). Analyses of national‐level data find that northern Brazil, which encompasses most of the Amazon Basin, posts the highest rate of infant growth faltering and poorest adherence to complementary feeding guidelines (UFRJ 2021, 2022; Lacerda et al. 2023). Yet only three studies of complementary feeding have been conducted in the north (Castro et al. 2009; Nogueira et al. 2022; Pizzatto et al. 2020). Findings include low dietary diversity (Castro et al. 2009), early introduction and high consumption of ultra‐processed foods (Nogueira et al. 2022), and inadequate maternal knowledge of WHO guidelines (Pizzatto et al. 2020). None of these studies attempted to identify the factors influencing the feeding decisions that resulted in the measured outcomes.

As the first study to explore complementary feeding in Belém, we had two aims: (1) to characterize the complementary diets of infants living in Belém and compare them to WHO recommendations and (2) identify the factors influencing mothers' feeding decisions during infants' first 2 years of life. To achieve these aims, we took a mixed methods approach that included data from 24‐h infant dietary recalls, a pile sorting activity, and in‐depth interviews which, together, allowed us to characterize infants' diets, and capture mothers' idealized patterns of food introduction and the factors influencing them.

2. Methods

2.1. Place and Sample

Belém (est. pop. 1.5 million) is the capital of the Brazilian state of Pará. The city, strategically located at the mouth of the Amazon River, has long been a center of commercial and cultural activity. As an economic hub, Belém is a site of continual within‐state rural–urban migration which contributes to a constant exchange of norms and traditions, including dietary practices (de Lima Amaral 2013). As a cultural center, it is recognized for its strong culinary identity, which is rooted in the integration of Amazonian and European foods and cooking traditions (de Sousa et al. 2020). Belém is also characterized by high rates of poverty and inequality, both the outcome of centuries of boom‐and‐bust economic cycles centered on the extraction of natural resources (Mansur et al. 2018; Rodrigues 2008). Despite some positive trends in infant nutrition at the national level between 2006 and 2019 (Castro et al. 2023), northern Brazil still posts some of the poorest infant/child health outcomes in the country. For example, the rate of stunting in the north is 8.7% while the national average is 7%. And, while nationally, 57% of infants meet MDD guidelines, the rate is only 44% in the north (UFRJ 2022) which may partially explain the higher rate of anemia in this region (17%) compared to the national average (10%) (Castro et al. 2023). Finally, the rate of infant overweight/obesity is on the rise nationally (6% in 2006 vs. 10% in 2019) with no differences between regions.

Poverty rates are also high in the north, with 46% of the population of Pará living below the poverty line and, in Belém, 21% live in conditions of extreme poverty (IBGE 2019). Within Belém, poverty is concentrated in low‐lying, densely occupied peripheral zones characterized by unplanned, precarious housing where residents often rely on clandestine access to municipal water and electricity, and where sanitation infrastructure is largely inadequate (Mansur et al. 2018; Rodrigues 2008). Wealthy residents, by contrast, occupy the city's core, where investment in infrastructure, including sanitation, roads, sidewalks, and parks, is evident, and government and commercial activities are concentrated. Within the core, wealthy residents commonly live in condominium complexes with pools and recreation areas (Gusmão and Soares 2018; Mansur et al. 2018).

We conducted this research as part of a larger, longitudinal study examining the sociocultural and ecological determinants of infant gut and oral microbiome development during the first 2 years of life. We recruited mother–infant dyads representing high socioeconomic status (HSES) and low socioeconomic status (LSES) households from private and public hospitals in the city's core, public health posts in two of the city's poorest neighborhoods, and via the Internet. The 30 mothers (10 HSES, 20 LSES) included here were drawn from the larger sample. To maintain participant anonymity, each mother was assigned a unique alpha‐numeric code that generally represents the location of her household in the city, although not necessarily the household's socioeconomic status. Our sample size is similar to related studies (e.g., Monterossa et al. 2012; Rodriguez‐Oliveros et al. 2014). In addition to the mothers, we interviewed four healthcare providers (three female, one male), three of whom were nurses working in public health clinics. The fourth ran a private physical therapy clinic focused on maternal health during pregnancy and the postpartum period.

2.2. Data Collection: Pile Sorting Activity and In‐Depth Interview

To address our study aims, we conducted in‐person interviews, in Portuguese, with mothers in their homes. The interview had two phases. In Phase I, we engaged mothers in a constrained pile‐sorting activity (Bennardo and de Munck 2014; Banna et al. 2015) where we asked them to sort a series of food cards into baskets representing the age at which they thought the food could appropriately be introduced into an infant's diet. While we presented mothers with age baskets as a guide (e.g., 1, 3, 6, 12‐months, etc.), mothers had the freedom to sort cards into infant age piles that fell between or even after those depicted on the baskets. To determine which foods to include, we used 24‐h dietary recall data collected from infants, one to 24‐months of age enrolled in the larger study and selected the most frequently reported items (> 20%). We shared this list with three key informants, all of whom were non‐participating mothers who lived in the study neighborhoods and asked for their input regarding the list's completeness. This two‐stage process resulted in 45 food cards. Table 1 organizes the cards into the food categories the WHO (2023) uses for calculating dietary diversity scores (i.e., MDD) for infants aged 6–23 months. Based on the Amazonian gastronomy literature (dos Santos and Pascoal 2013) and feedback from key informants, we grouped some food items and presented them on the same card. In Table 1, a shared superscript denotes these groupings. Fruits serve as an example. The Amazon is known for its rich diversity of fruits (endemic/introduced, domesticated, managed, wild) (Cavalcanti 1972). Residents of the region tend to classify foods, including fruits, into categories based on their perceived properties (Maués and Maués 1978). This ethnomedical system, referred to as reima, is grounded in Hippocratic‐Galenic humoral theory introduced into the region during Portuguese colonization in the 16th century (Foster 1994; Prado et al. 2022). Like the hot/cold classification system documented across Latin America (Currier 1966; Messer 1981), reima emphasizes the need to maintain balance in the body to ensure wellbeing. Both systems recognize that ingested items can affect this balance. However, instead of hot/cold, reima classifies foods as reimoso, items that can prejudice health, and manso, those that are considered benign or safe to eat. Reimoso foods tend to be fatty, acidic, strong in flavor, or derived from animals seen as aggressive or unclean.

TABLE 1.

Food cards organized by WHO food categories used to calculate infant (6–23 months) dietary diversity.

WHO food categories	Infant food card
Category 1: Breast milk	Breast milk
Category 2: Grains, roots, tubers, and plantains	Rice, farinha (a toasted meal made from the bitter manioc root ( Manihot esculenta Cranz)), farofa (farinha cooked with butter/onion), tapioca (starch extracted during farinha production), mingau (porridge made from ultra‐processed mixes with a base of rice, wheat, or corn), mingau caribé (traditional porridge made from farinha), bread, pasta, tucupi (liquid condiment made during farinha processing), crackers, cookies, cake (wheat, manioc, tapioca)
Category 3: Pulses, nuts, and seeds	Beans, lentils, a chickpeas, a soybeans a
Category 4: Dairy products	Milk (liquid/powdered), b composto (ultra‐processed, fortified and sweetened powdered milk byproduct), b ultra‐processed flavored‐yogurts, infant formula
Category 5: Flesh foods	Chicken, beef, pork, offal, scale fish, skin fish, shrimp, c crab c
Category 6: Eggs	Eggs
Category 7: Vitamin‐A rich fruits and vegetables	Mango, d papaya, d passion fruit, e taperebá ( Spondias mombin ), f broccoli, g kale, g cabbage, g carrot, g sweet potato, g natural juices from these fruits
Category 8: Other fruits and vegetables	Banana, d red and green grapes, d apple, d pear, d melon, d watermelon, d plum, d orange, e tangerine, e avocado, e Brazil nut, e pineapple, e strawberry, e cupuaçu ( Theobroma grandiflorum ), f bacuri ( Platonia insignis ), f murici ( Byrsonima crassifolia ), f beet, h zucchini, h chayote ( Sechium edule ), h potato, h açaí ( Euterpe oleracea ), jambu (Acmella oleracea), natural juices from these fruits
Category 9: Other local foods	Water, coconut water, coffee, soup (vegetables, mocotó, noodles), tea, broth (bean, chicken, fish), jambú, soft drinks, bonbons, i lollipop, i artificial juices (powdered mix), chocolate, j chocolate milk powder, j industrially produced baby food

^a Lentils, chickpeas, and soybeans were combined.

^b Powdered milk and composto were combined as many mothers used them interchangeably as they were unable to distinguish between them due to similarities in packaging.

^c Shrimp and crab were combined.

^d Fruits I: mango, papaya, banana, grapes, apple, pear, melon, watermelon, and plum.

^e Fruits II: orange, tangerine, avocado, Brazil nut, pineapple, and strawberry.

^f Fruits III: passion fruit, taperebá, cupuaçu, bacuri, murici.

^g Vegetables I: leafy and green vegetables locally referred to as verduras were combined.

^h Vegetables II: beets, zucchini, chayote, potato are locally referred to as legumes and were combined.

ⁱ Candies: bonbons and lollipops were combined.

^j Chocolate and chocolate milk powder were combined.

Guided by our key informants, we adopted the local classification system and divided the fruits into three groups. Fruits I consists of domesticated fruits (endemic and introduced) perceived to be of low acidity/fattiness. Fruits II includes fruits considered moderately acidic or fatty. Fruits III includes managed/wild species considered highly acidic or fatty. Açaí, an endemic managed palm fruit, was presented on its own card as it is a local dietary staple. In the text and table, we translate food items into English unless there is no English equivalent, in which case we use the Portuguese word and provide a short description and/or the scientific name.

Once the mothers completed the pile sorting activity, we initiated Phase II. Beginning with the first age pile and continuing through the last, we noted which cards were placed in each pile and then asked mothers to explain their rationale for placing those foods in that pile. As mothers explained their decisions, the conversation evolved, allowing us to follow up mothers' responses with additional questions to clarify ideas and ensure understanding. Interviews with the four healthcare providers focused on the types of health services mothers had access to, current feeding recommendations, and their perspectives on complementary feeding practices among the populations they serve. We audio recorded all interviews.

2.3. Maternal and Household Characteristics

We used structured interviews to gather data on maternal and household characteristics shown to be associated with infant feeding decisions. For the mothers, this included age, parity, educational attainment, and employment status (Mosquera et al. 2019). Educational levels in Brazil are as follows: Primary I (Grades 1–5), Primary II (Grades 6–9), secondary (Grades 10–12) (IBGE 2019). Higher education is any postsecondary education. We characterized maternal employment status as formal (signed labor booklet) and informal (not signed). In terms of household characteristics, we included demography (number of permanent residents), income, and hired help (maids and nannies). To classify household SES, we used the criterion established by the Brazilian government to assess eligibility for social support programs, whereby a household is considered LSES if total earnings are less than three times the minimum wage (Caixa Econômica Federal, n.d.).

2.4. Data Analysis

We used descriptive statistics to characterize the sample. To address aim one, we used data from the pile sort activity. First, we calculated the frequency each of the 45 food cards (derived from 24‐h dietary recalls) were placed in the different infant‐age piles. This provided two pieces of information: (a) a general pattern of food introduction and (b) the level of agreement among mothers regarding the timing of introduction. For example, all mothers placed breastmilk in the one‐month pile; thus, there was 100% agreement. For foods placed in piles that fell between seven and 12 months, we combined them into one group (7–12 months). We also combined foods placed in age piles later than 15 months or in a pile indicating the mother had no intention of feeding the food to her infant into a single group (later/never). As studies have identified household socioeconomic status as a factor influencing maternal feeding practices, we used chi‐squared tests to identify differences in the sorting of the 45 cards between mothers of HSES and LSES. As none of these were significant, we discuss findings for the group (n = 30). We used the NOVA classification system to categorize foods/beverages by level of processing (i.e., unprocessed/minimally processed, processed culinary ingredients, processed foods, and ultra‐processed foods) (Martinez‐Steele et al. 2023).

To address aim two, we transcribed each audio file. Each transcript was reviewed and corrected by three of us (A.C.B.A., B.A.P., T.C.). Once the files were clean, we read through them multiple times, with a focus on how mothers explained their sorting decisions. Based on these notes, we developed a code book. BP conducted the initial coding of the transcripts using MAXQDA 24 software, which was refined through checking by and conversations with the other authors. Once coding was complete, we reviewed the coded material and met to identify and discuss emergent themes following standard qualitative analytical practices (Graneheim and Lundman 2004). Review of the qualitative data indicated we reached saturation (Fusch and Ness 2015) and, thus, had a sufficient sample size to explore the topic of complementary feeding.

3. Results

3.1. Characteristics of the Sample

Table 2 reports the sociodemographic characteristics of the women who participated in the study. Average maternal age was 32 years and ranged from 20 to 44; 50% were first‐time mothers. In terms of educational attainment, 20% of mothers had completed only a primary education, 53% a secondary education, and 27% had taken college‐level courses or completed a college degree. Less than half (40%) of the mothers were employed; 92% of whom worked in the formal sector. Average household size was 4.1 and ranged between two and eight. Of the 30 households, six (20%) had domestic workers; all but one of which was classified as HSES.

TABLE 2.

Sociodemographic characteristics of the n = 30 participating mothers.

	Total (%)	Mean ± SD
Mothers age (years)	30 (100)	31.6 ± 6.4
20–27	10 (33.3)	24.6 ± 2.3
28–35	10 (33.3)	31.5 ± 2.8
36–44	10 (33.3)	38.6 ± 2.9
Parity
1	15 (50.0)
2	11 (36.7)
3+	4 (13.3)
Education
Primary I/II	6 (20.0)
Secondary	16 (53.3)
Higher education	8 (26.7)
Employment
Unemployed	18 (60.0)
Employed	12 (40.0)
Formal sector	11 (91.7)
Informal sector	1 (8.3)
Household SES
HSES	10 (33.3)
LSES	20 (66.6)
Household size (all)	30 (100)	4.1 ± 1.3
2–3 members	11 (36.6)
4–5 members	15 (50.0)
6–8 members	4 (13.3)
Domestic workers in home	6 (20.0)

3.2. Pile Sort: Characterizing the Complementary Diet

Figure 1 reports the findings of the pile sort activity, with the 45 food cards listed in order of timing of ideal introduction into infants' diets, from earliest (1 month) to latest (later/never). The cells are shaded, darkest to lightest, based on the percent of mothers who placed the food card in that age pile; the number in the cell is the percentage. An unshaded cell indicates no mothers placed the food card in that pile.

FIGURE 1.

Results of the pile sort activity.

While we sorted the foods by mothers' idealized timing of introduction, to facilitate discussion here, we organize them into five groups (1, 2a, 3, 2b, and 4). The cut‐offs and titles of the groups are based on shifts in timing of introduction and local food classifications as revealed in the ethnographic interviews.

Group 1 consists exclusively of liquids. Except for breastmilk, there was limited agreement as to when liquids should be introduced into infants' diets. Apart from infant formula, an ultra‐processed food, these liquids are un/minimally processed. Group 2a (baby's first menu) includes a range of fruits and their juices, vegetables, often served in soups, as well as mingau, fish (w/scales) and rice. Except for mingau, most mothers agreed that 6 months is an appropriate time to begin introducing solid foods and that these foods are most suitable. All except mingau, an ultra‐processed food that, depending on the brand, can contain added vitamins and minerals, are un/minimally processed. Group 3 (soft/convenience foods) consists of items mothers considered snacks. They range from un/minimally processed foods such as Fruit III and caribé to ultra‐processed items such as commercial, ultra‐processed baby food, crackers, cookies, and artificially flavored yogurts, some of which contain added vitamins and minerals. While a small percentage of mothers placed these foods in the 3‐month pile, thus their position in the table, the majority placed them in either the 15‐month or later/never piles. As with Group 1 (liquids), there was limited agreement among mothers as to when an infant could consume these foods. Group 2b (extended first menu) includes several animal‐based foods and additional dietary staples. What distinguishes Group 2b from 2a is that while many mothers also placed these items in the 6‐month pile, in cases where they did not, they were placed in later age piles. While there was relatively high agreement among mothers that beans/broth, pasta, beef, eggs, and chicken could be introduced at 6 months and that tapioca is best introduced a bit later, there was less agreement as to when the other items in this group could be given. Foods in this group are un/minimally processed except for composto. The final group, Group 4 (later/never), includes foods that most mothers agreed should be given later in life or not at all. As seen in Figure 1, there was significant agreement among mothers that ultra‐processed foods such as candy and chocolate (solid and powdered), soda/artificial juice (powdered mix), and cake were generally not appropriate for infants. There was less consensus regarding the un/minimally processed items such as jambu, farinha/farofa, pork, and shrimp/crab.

In summary, the pile sorting exercise revealed a model of complementary feeding that, except for some liquids and fruits, begins around 6 months with foods from category 2a and evolves over time to include foods common in the regional diet (categories 2a + 2b). If we consider Groups 2a and 2b together, we find that the ideal complementary diet during the first year includes foods representing all nine WHO food categories (Table 1). While the pile sorting activity helped characterize an ideal trajectory of complementary feeding in this urban Amazonian setting, the accompanying interviews allowed us to explore the factors influencing this trajectory and feeding decisions.

3.3. Biocultural Model of Complementary Feeding in Urban Amazonia

Analyzing the n = 30 interviews, we identified three themes and nine categories (Figure 2), which together indicate that for these mothers, complementary feeding is guided by the convergence of knowledge systems that inform perceptions of food/eating and infant development/health and is implemented within a socioeconomic and cultural context that can support and constrain its daily practice.

FIGURE 2.

Emergent themes, categories, and codes from the n = 30 interviews.

3.4. Theme 1: Integrating and Applying Trusted Sources of Advice

Theme 1 one contained three categories, the biomedical healthcare system, Internet/social media, and family and friends. Regardless of economic circumstances, cultural expectations are that mothers are ultimately responsible for physically feeding infants, as well as deciding feeding schedules, and the composition of infants' diets. To make these decisions, mothers reported relying on advice from what they perceived as trustworthy sources, some of which they actively sought out. Of the three major sources of advice, healthcare providers proved, particularly, influential. Mother RM‐403 told us “I prefer to follow the doctor … Doctors have the knowledge, you know, they studied more.” The healthcare providers we interviewed confirmed that mothers in Belém can schedule consultations with nurses, pediatricians, nutritionists, and other specialists; although, the latter two are more often accessed by those with private health insurance. These providers are trained to advise mothers on infant nutrition and often provide them with “menus” derived from the Brazilian infant feeding guide. A common refrain we heard from mothers when discussing foods in Groups 2a and 2b (Figure 1) was that a doctor/nurse/nutritionist told them it was the appropriate time to introduce it. Mother AR‐161 told us “She [doctor] gave me a menu with some fruits … she put banana, papaya, and pear. But she told me I could give her mangos …” Another mother (G‐208) said “The nutritionist said that I could start with everything … fruits, vegetables and also fiber and meat, fish, all types of meat at six months.” The placement of certain foods in the 15‐month or later/never piles was also influenced by healthcare provider advice to avoid sugar, salt, and ultra‐processed foods, which mothers explained was to avoid future health problems. Mother RM‐445 told us, “Sweets, too. Soda, I know it's not good. I'm trying to avoid it precisely to prevent diseases in the future, right? I imagine that the longer I delay the introduction of these sweet, processed foods, I'm postponing illnesses too.”

While all mothers reported getting in‐person advice from healthcare providers, many also sought information on the Internet. As mothers understood, on‐line forums focused on infant care and feeding were developed by doctors/nutritionists, which allowed them to readily access what they viewed as reliable and up‐to‐date advice. Mother AR‐170 told us “I searched the Internet for information on how to feed him … from some websites that claim to be specialists in infant nutrition” while Mother AR‐111 said “I downloaded an app called BLW (Baby Led Weaning) Brazil and relied on it for recipes and menus, you know, so that each day there were different foods.” Mothers also accessed social support on‐line. Mother AR‐141 said “I have a group of women who had babies around the same time as me. In this group there are 150 women, and we exchange ideas, experiences.”

While the influence of healthcare professionals is undeniable, mothers also discussed the role female kin, particularly their mothers, mothers‐in‐law, and grandmothers, and friends played in their feeding decisions. However, there was a notable difference between these sources of advice. While healthcare professionals provided informational support in the form of nutritional recommendations, family and friends were in the home, where they additionally provided instrumental and emotional support. Mother AR‐100 recounted how the pediatrician recommended feeding vegetables but not how to prepare them to retain their nutritional value. She said “He didn't explain that if you cook vegetables too much, they lose all their nutritional value. So, what helped me with this part was my mother. She even gave me the pot [double boiler] that we still use today.” She went on to explain how using the double boiler allowed her to add the liquid and the nutrients it contained back to the vegetables before feeding them to her daughter. Mother TF‐321 described how her mother taught her to mash fruits and vegetables. Mother AR‐114 said “On a daily basis … I usually consult my mother‐in‐law … who lives with me … I ask her, this food, do you think it is good? … I end up consulting her a lot.” Mother RM‐425 recalled how her friend advised her on food preparation “I went to a friend of mine who has more experience, has four children, is 40 years old [who told me] ‘no, make it without sugar, without milk, the more natural, the better.’ Advice from those with more experience. I think it's better.”

Overall, mothers indicated receiving consistent advice across trustworthy sources regarding the timing of introduction of many of the foods in Groups 2a and 2b (Figure 1) as well as the ultra‐processed items in Group 4. This consistency helps explain high agreement among mothers in sorting these food cards. However, there were exceptions where advice of family and health professionals differed. In these cases, mothers had to manage daily pressure/interference from relatives and potential admonishment from healthcare providers (see Theme 3). Contradicting advice generated doubt and helps explain observed lack of consensus (Figure 1). For example, while public health professionals emphasize that liquids other than breastmilk are not needed prior to 6 months, intergenerational advice included giving water and coconut water as early as 1 month in response to the year‐round intense heat in the city, and herbal teas to treat illness. Mother RM‐423 told us “Tea is for the gas caused by colic. You understand? They [elders] say it's good.”

3.5. Theme 2: Infant Readiness and Future Health

Theme 2 contains three categories, food characteristics, infant development, and size. In northern Brazil the three intersect as part of the reima ethnomedical system. The reima system integrates an individual's bodily state and the characteristics of foods in making dietary pre‐ and proscriptions. A body in an unbalanced or vulnerable state, for example, the immediate postpartum or postoperative, cannot handle foods classified as strong or irritating due to their acidity, fattiness, or ability to induce reactions, including allergic reactions. Thus, people in these states must avoid such foods. Mother G‐251 told us “In the past, people said [a food] was bad. This food here, it is reimoso. If the person has a problem, a little wound, it is harmful. You still hear this today. They don't eat that.” Small size and immaturity make infants vulnerable, which was articulated by mothers using diminutive terms, for example, menorizinha (being little or small), estômagozinho (little stomach). In terms of foods from Figure 1, Fruits III (strong/acidic), skin fish (strong/irritating), jambú (strong), pork (strong/fatty), shrimp/crab (strong/irritating), and tucupi (strong/acidic) are commonly classified as reimoso. While the healthcare professionals we spoke with largely rejected pre‐ and proscriptions of the reima system based on their biomedical training, as members of the society, they understood them, recognized that mothers' feeding decisions were influenced by them, and made efforts to accommodate them when giving feeding advice by offering alternatives. One of the female nurses told us “I tell her [the mother], but if you're afraid, let's respect it, because we also can't clash with the person's culture, otherwise they'll avoid us [healthcare system] because they won't gain trust.”

A rich discussion with mother RM‐403 illustrates how reima influences infant feeding decisions. Discussing Fruits III she told us “I think these are very, very strong to give to a one‐year‐old baby, passion fruit, cupuaçu, taperebá, bacuri, murici, I think they are too heavy to give now, but as she grows, we can let her try.” She further explained “They can upset her little stomach, because she's smaller, because she's still a young child … So, it's better you avoid it.” Talking about pork, she said “it's very reimoso and can be bad for her stomach … I believe when she is big, bigger, eight, ten years old.” She felt the same about skin fish, shrimp, and crab. Mother G‐223 related the health risks to her son's developmental vulnerability, “shrimp is also a little reimoso, as it can suddenly cause diarrhea. His teeth are coming in, and it could be bad for him.” Tucupi and jambú were discussed in a similar way. Overall, concerns about how the characteristics of reimoso foods interact with infants' small and immature bodies led many mothers to place these items in later age piles. This contrasts with how mothers talked about the foods from Groups 2a and 2b (Figure 1). Mother G‐205, in discussing Fruits I said, “because these are lighter to start with … less aggressive, right?” Mother RM‐403 told us “…egg is light, egg is good and calm (i.e., non‐irritating).” Mother G‐223 expressed these sentiments about additional items largely placed in the 6‐month pile “Meat, soups, they are light, fish.”

Markers of infant development increased mothers' willingness to introduce foods with stronger characteristics or perceived to be harder to process and/or digest (i.e., foods from the latter part of Group 2b and initial part of Group 4). Mother AR‐100 explained “from what I understand, this is when the digestive system will be somewhat mature enough to digest, right? And, to metabolize the compounds, right? The baby's liver, kidney … because the baby has to be a little strong to swallow, right? His spine, to support himself in a way that he can swallow well.” The ability to swallow influenced mothers' willingness to introduce several staple carbohydrates including rice, farinha, farofa, and tapioca, and some legumes, which they felt were choking hazards. The presence of teeth seemed to be particularly influential, especially in the timing of introduction of meat/offal and legumes, which, unlike fruits and vegetables, mothers tended not to mash. Mother G‐261 in discussing beef and chicken said “I think nine months, because it is a more rigid meat. So, she will need to chew.” Mother AR‐161 told us “Lentil/garbanzo beans, I believe at nine months he should have some of his teeth formed to be able to chew these.”

While size, as a marker of development, encouraged the introduction of new foods, as a marker of health, it signaled different things, largely based on one's generation. As mother AR‐161 informed us “In the past, a pretty baby was a fat baby. Now we know that's not the case, right? I believe the culture is changing. People are starting to realize that diseases that people have could be prevented in childhood. Avoiding obesity means controlling a baby's diet, encouraging him to follow a healthier eating pattern.” Mother G‐202 recounted a conversation with the pediatrician who told her “Mom, she's obese … stop giving her mingau, Danone (flavored yogurt), instant soups (ramen).” While public health messaging and interactions with healthcare providers led many mothers to express concerns over infant fatness, older generations tended to have a different point of view (Theme 1). Mother RM‐423 told us that if it were up to her mother, she would be feeding mingau, a pap used to supplement or even replace breastmilk, within the first month; “from the beginning, she [mother] did not think my breastmilk was enough. But the pediatrician said ‘no, it is enough, it is good.’” Mother G‐261 recounted “my grandmother said I should prepare milk [composto], give mingau, all those kinds of foods … my nutritionist [upon finding out] gave me a firm warning.” Attempting to understand intergenerational differences regarding optimal infant body size and fatness, mother AR‐161 did some research.

We were unsure as a lot of people talk about mingau, instant foods. So, I did some research, its composition is all flour. It's just to make the child gain weight. So, I started researching why the pharmaceutical industry still releases this type of food, claiming it to be healthy for children. I discovered that in the past there were many malnourished children … So, the pharmaceutical industry joined forces with the food industry and established these foods to help children gain weight. But they remain today, right … and they [mothers] stop feeding healthy things, because these are quicker … sometimes the quickest thing is not so healthy. That's why I personally do not offer them.

Many mothers viewed commercial baby food, carbohydrate‐rich foods, for example, farinha/farofa, bread, and mingau, and ultra‐processed sweets, for example, candy/chocolate, and soda, similarly, which helps explain why many placed them in later age piles or indicated never intending to feed them to their baby (Figure 1).

3.6. Theme 3: Challenges Realizing Feeding Ideals

Theme 3 three contains three categories, cost, interference of others, and mothers' work, all of which challenged mothers' ability to feed infants as they desired. Most notable were financial constraints, the impact of which manifested in different ways. For some it was simply not having the funds to buy the variety of recommended foods. Mother G‐202 told us “Sometimes yes, sometimes no [regarding money to buy recommended foods] because in this house everyone [except her] is unemployed.” For others it was fear of wasting limited resources on a new food the child might reject, since many of the recommended first foods (i.e., fruits/vegetables) are not part of the family diet, especially in low‐income homes. Mother TF‐321 said “So if I make food separately [for the infant and she does not accept it], I waste it. So, to avoid wasting it, I make foods that everyone eats.”

When opinions and advice from healthcare providers and family members were incongruent, some mothers reported relatives pressuring them or, even, feeding their infants foods they did not want them to have. Mother RM‐445 recounted the pressure she received from family during a trip “There everyone was saying ‘He does not eat that? Why not? Poor thing!’ You know? It is a struggle when you vary your routine.” Mother G‐261 told us “It is unnecessary to give crab to a baby. But she has already eaten it against my will … because I sometimes leave her at my mom's house and there, they give her everything.” Sweets and snacks (e.g., crackers/cookies) are also commonly offered to infants as a treat by others in the home, sometimes without mothers' permission.

A final challenge was work. In this sample, 40% of mothers worked outside the home, mostly in the formal sector. Aware of the need to return to work, mothers reported preparing their infants for their absence by introducing foods prior to 6 months and healthcare providers offered advice on what foods to start with (e.g., Group 2a foods, formula). Mother G‐261 told us “It was necessary because my husband was unemployed, and I work … I had to keep her fed so I could work. So, I had to, one way or another, introduce foods into her diet.” Mother G‐205 told us “Since he was five months old, he was drinking powdered milk/mingau. I couldn't wait because I needed to go back to work … and I couldn't afford to buy NAN 1 [brand of formula].” Due to time constraints, mother G‐202 reported giving her child convenience foods, specifically flavored yogurt, crackers, and instant soups, due to their ease of preparation “they are so simple to make, so I was giving them.”

4. Discussion

Previous work has identified a range of socioeconomic (e.g., Cavalcanti and Boccolini 2022; Gatica‐Domínguez et al. 2021; Maciel et al. 2018) and cultural factors (e.g., Dutta et al. 2006; Monterossa et al. 2012; Rodriguez‐Oliveros et al. 2014; Zobrist et al. 2017) influencing complementary feeding, with strong emphasis on the timing of initiation. Despite calls for a more holistic approach (e.g., Engle et al. 1999; Macadam and Dettwyler 1995; Pelto et al. 2003; Popkin 1986), less consideration continues to be given to how these factors intersect to influence complementary feeding decisions and how they evolve as the infant develops. Integrating data from a pile sort activity and in‐depth discussions, we offer a biocultural model that illustrates how macro, intermediate, and individual‐level factors intersect to shape complementary feeding in Belém, Brazil (Figure 3). The model can serve as a framework for studying complementary feeding and designing interventions across contexts.

FIGURE 3.

Biocultural model of complementary feeding.

4.1. Pathway 1: Direct Influences of the Political‐Economic and Cultural Contexts on Feeding Decisions

Pathway 1 (Figure 3) depicts the direct influence interactions between the political‐economic and cultural contexts have on infant feeding through access to time, money, and power. As is the case globally (https://genderdata.worldbank.org, 2024), in Belém there is great demand for women's labor, in and outside the home, unpaid and paid. Recent data show that in northern Brazil, the average weekly number of hours spent on unpaid house and care work is almost twice as high for women as it is for men (20.3 and 11.5 h, respectively) (IBGE 2022). For mothers employed in the formal sector, Brazilian law guarantees a minimum 4‐month maternity leave; fathers get 5 days (Brasil Ministério da Saúde Secretaria de Atenção à Saúde Departamento de Ações Programáticas Estratégicas 2017). For those working in the informal economy, there are no protections. Such policies reflect and reinforce the norm that women are responsible for infant care, which includes meal planning, shopping, preparation, and physical feeding. At the same time, these policies can undermine mothers' feeding ideals. As with other Brazilian‐based studies (Boccoloni et al., Boccolini et al. 2015; Scavenius et al. 2007), we found the need to return to work influenced the duration of exclusive breastfeeding. Its impact on complementary feeding, especially the adoption of easy‐to‐prepare, ultra‐processed foods, is less well‐documented. Like da Gloria Miotto Wright et al. (1989), we found the return to paid labor and gender inequalities in time dedicated to unpaid childcare and housework influenced mothers' decisions to incorporate ultra‐processed foods into infants' diets. While studies indicate a positive effect of maternal employment on infant dietary diversity, they rarely control for ultra‐processed foods in these calculations (Oddo and Ickes 2018). This is problematic due to concern over how early and continued exposure to ultra‐processed foods affects short‐ and long‐term health (Baker et al. 2021, 2023; Oliveira et al. 2022; Pérez‐Escamilla et al. 2023).

Intersections between the political, economic, and cultural contexts also directly influence access to support for caring for infants and control over that care. Wealthier mothers can offset childcare demands by hiring other women—maids and nannies. In Brazil, women make up over 90% of domestic workers, and national data illustrate their impact. Women in the highest income quintile dedicated 7.3 fewer hours per week to unpaid domestic work than women in the lowest quintile (IBGE 2022). In Belém, maids and nannies shopped for and converted fresh produce and meats into infant‐friendly dishes, fed infants, and cleaned up afterwards. Unable to afford this help, lower‐income mothers relied on publicly provided daycare or the unpaid labor of family, which left them with less control over what their infants were fed. They had no power to dictate daycare practices, and when the beliefs and practices of family members, whose support they relied on, conflicted with their ideals, they were in no position to challenge them—a scenario also documented in other settings (e.g., Cheney et al. 2019; Jerónimo et al. 2021).

Economic access to food directly influenced the composition of the complementary diet, a global phenomenon (Cavalcanti and Boccolini 2022; Gatica‐Domínguez et al. 2021; Herman et al. 2023; Nogueira et al. 2022). The ability to afford not only the variety of recommended foods but also the risk of the infant rejecting them and them going to waste affected low‐income mothers' feeding decisions. Across contexts, dietary staples rarely include the range of foods recommended by the WHO (2023). In low‐income households in Belém, fruits and vegetables are not part of the dietary repertoire. This contrasts with high‐income households where staples are accompanied by vegetable dishes, and fruit is consumed as snacks and dessert. Thus, high‐income mothers are better able to provide their infants opportunities to experiment with and integrate new healthy foods into their diets. Overall, in Belém, as elsewhere, the political‐economic context influences both the timing of initiation of complementary feeding and the composition of the diet—with potential long‐term effects on infant health.

4.2. Pathway 2a/2b: The Integration of Ethnomedical Systems and Subsequent Food Choices

Authoritative knowledge regarding complementary feeding came from two dominant sources (Figure 3). Embedded in the broader political–economic context is the biomedical healthcare system, which disseminates global public health guidelines. Mothers accessed this knowledge through formal education, the Internet, and uptake of healthcare, largely without participation from fathers or other family members. Within the cultural context is the inter‐ and intragenerational circulation of knowledge and wisdom gained from life experience, which, in Belém, included the rules and application of reima. Mothers accessed this knowledge through their female social networks, again with little male involvement.

Low maternal educational attainment, and beliefs about food and infant development that conflict with biomedical recommendations are cited as barriers to a nutritionally adequate complementary diet (Alvarado et al. 2005; Dutta et al. 2006; Monterossa et al. 2012; Pizzatto et al. 2020; Rodriguez‐Oliveros et al. 2014; Zobrist et al. 2017). We found limited evidence that these factors compromised infant feeding in this setting. All mothers were aware of the recommendation to exclusively breastfeed for 6 months and expressed a desire to follow it. Furthermore, mothers indicated that a large range of the foods present in infants' diets were appropriate to introduce as early as 6 months, meaning they could theoretically meet the WHO (2023) guidelines for a diverse complementary diet. Thus, while differences in food classification and thus their appropriateness for infant diets existed between the biomedical and reima systems, mothers could reconcile them in ways that would not compromise infant nutrition. We found some evidence that this might be related to the efforts of healthcare providers. Primary healthcare providers in Brazil are trained to give mothers infant feeding advice (Brasil Ministério da Saúde Secretaria de Atenção à Saúde Departamento de Ações Programáticas Estratégicas 2017) and we found that among the four we spoke with, there was a willingness to harness their knowledge of both ethnomedical systems to aid mothers make healthy feeding decisions.

The reima system has little to say about ultra‐processed foods newer to Brazilian infants' diets and not easily characterized as reimoso or manso. Thus, mothers' opinions of these foods are largely derived from the biomedical system. Mothers in this study were both aware and had a negative opinion of ultra‐processed foods, a finding that contrasts with Pizzatto et al. (2020) in their study of complementary feeding in Maranhão, Brazil. Yet, despite their knowledge, mothers in Belém still offered ultra‐processed foods to their babies. Following the International Code of Marketing of Breast‐Milk Substitutes (WHO 1981), Brazilian law regulates the advertising and labeling of foods marketed at young children (Brasil Ministério da Saúde Secretaria de Atenção à Saúde Departamento de Ações Programáticas Estratégicas 2017). These laws, however, are principally aimed at encouraging breastfeeding. There is little regulation of ultra‐processed foods marketed at those > 6‐months, including commercial mingau, crackers, artificially flavored yogurt, cookies, and composto. In northern Brazil, the traditional mingau caribe was made by extracting the starch from farinha and adding sugar and sometimes powdered milk. Today mingau caribe is largely replaced with mingau made from industrialized powdered mixes that have a base carbohydrate such as rice, wheat, or corn (common brands are Cremogema [Unilever] and Mucilon [Nestlé]) and include artificial colorings, flavorings, and fillers. While some are fortified, all are part of a repertoire of ultra‐processed foods marketed toward infants.

Recent data documents a surge in availability of these foods in the Global South, including Brazil; the result of lobbying efforts and legal challenges to government regulatory action by the baby food industry (Baker et al. 2021). High availability, economic accessibility, and convenience all contribute to their utilization (Louzada et al. 2023; Monteiro et al. 2011; Soares et al. 2021; United Nations Children's Fund [UNICEF] 2020). Lacerda et al. (2023) found that over 85% of infants (6–24 months) living in northern Brazil were receiving ultra‐processed foods. Among infants in southern Brazil, Spaniol et al. (2021) found ultra‐processed foods contributed as much as 33% of all calories consumed.

4.3. Pathway 2b/3: How Integrated Ethnomedical Systems Influence the Interpretation of and Response to Infant Biobehavioral Cues

We found important differences between the two ethnomedical systems in terms of how biobehavioral cues guide feeding decisions. Biomedical models largely use age as a proxy for infant development and to guide feeding recommendations (WHO 2023), making them less responsive to individual variation. In contrast, the reima system prioritizes balance and emphasizes individual needs. Similar to other studies (e.g., Matvienko‐Sikar et al. 2018; Rodriguez‐Oliveros et al. 2014; Zobrist et al. 2017), we found mothers assessed individual needs/readiness via markers of development, that is, head and tongue control, ability to sit upright, emergence of teeth, and used them to guide the introduction of new foods. In cases of precocial development, mothers modified biomedical advice to accommodate their infant's needs.

A second difference is the interpretation of infant weight and fatness. National‐level data document significant increases in overweight and obesity among Brazilian adults (IBGE 2022) and young children (Castro et al. 2023; UFRJ 2022) since the early 2000's. In response, public health campaigns are increasingly promoting healthy body size and related eating habits, including for children. Mothers in Belém struggled to reconcile this biomedical advice with longstanding views that chubbiness is a sign of health, a viewpoint reinforced by older relatives (see also Lindsay et al. 2009). Mothers reported that older relatives interpreted infant thinness and crying as signs of breastmilk insufficiency and, in response, encouraged early supplementation (e.g., commercial mingau and composto). This finding is consistent with that of Scavenius et al. (2007) who reports a high rate of perceived milk insufficiency in Brazil and discusses its role in shaping complementary feeding patterns. These intergenerational differences, which were a source of strife in some households (see also Jerónimo et al. 2021), may be related to older women's direct experiences rearing children during the 1970–1980s when the rate of infant malnutrition in Brazil, especially in the north and northeast, was high (Escoda 2002; Lindsay et al. 2009; Oliveira 2004). As discussed, while infant nutritional status has improved nationally, the north still lags other regions in terms of stunting and anemia, although not for weight‐for‐height or overweight/obesity (UFRJ 2021). Thus, the situation in the north appears to be an emerging dual burden, where infant/child under‐ and over‐nutrition exist within the same population or, even, individual (Tzioumis and Adair 2014). Some attribute this trend to the use of ultra‐processed foods (Lacerda et al. 2023; Soares et al. 2021) which serve as no‐prep substitutes for traditional items (e.g., commercial mingau replaces traditional mingau caribé). Both characteristics contribute to their adoption—a familiar food among time‐constrained caregivers. While some ultra‐processed foods are fortified, they are also heavily marketed, relatively cheap, highly palatable, and calorific, which can encourage overconsumption and weight gain. It is within this changing and complex context that mothers in Belém are making feeding decisions.

Taken together, the data suggest that targeting maternal education alone will not improve complementary feeding. Instead, efforts should address the context in which feeding occurs—including demands on mothers' time that result from the gendered division of labor and differences in intergenerational beliefs and power dynamics—as both limit mothers' agency in exercising their knowledge and implementing their feeding ideals (Jerónimo et al. 2021; Cheney et al. 2019). Additionally, the importance of complementary foods and feeding practices for infant growth and development during the first 2–3 years of life must be prioritized in policies that govern the marketing and distribution of ultra‐processed foods (Baker et al. 2021).

5. Conclusion and Future Directions

Our findings and the biocultural model they inspired emphasize the need to understand complementary feeding decisions within the broader political, economic, and cultural contexts in which they occur. We underscore how the intersection of these contexts continues to hold mothers responsible for complementary feeding but not necessarily empower them to make decisions or support their efforts. This is especially the case among women of lower socio‐economic status who face greater challenges accessing foods required for an adequate complementary diet and who depend on intergenerational support for infant care. This is especially challenging and stressful as the food environment and biomedical advice (diet, infant body size norms) evolve. As key sources of authoritative knowledge, our data suggest that healthcare professionals can bridge conflicting advice, especially if messaging were to target other influential individuals in mothers' lives who can share in the responsibility of infant feeding. Future research with a larger number of providers from the public and private health sectors is required to explore this topic. As political–economic shifts across Latin American have altered household economics and gender roles (Chant 2002) future research on the role of fathers in complementary feeding is warranted. While our in‐depth, largely qualitative exploration of the biocultural factors influencing mothers' complementary feeding decisions in the urban Amazon is not generalizable, the model we developed from these data integrates many of the factors mentioned in other studies. Thus, it should prove useful in guiding new and needed research across a broad range of settings.

Author Contributions

B. A. Piperata: conceptualization, methodology, formal analysis, writing – original draft, review and editing (lead), funding. S. F. Fannin: conceptualization, methodology, data collection, formal analysis, review and editing. T. Cestonaro: formal analysis, review and editing. A. C. Brito‐Azevedo: formal analysis, review and editing. V. de Cássia Tavares da Silva: methodology, review and editing. J. Mendonça Freire Pereira: review and editing. R. Bittencourt Tavares Oliveira: review and editing.

Ethics Statement

The study received institutional review board approval from the Comissão Nacional de Ética em Pesquisa (CONEP) (número de parecer 4.043.620), the Brazilian federal agency that reviews international human subjects research, and the Ohio State University (2020B0014). At the time of recruitment, members of the field team provided an overall description of the project and shared a copy of the informed consent form with the potential participant. A team member then reviewed the form verbally, stopping to answer questions as needed. Upon agreement, participants signed two copies, one of which they kept. We retained the other copy.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.