Using formative research to design context‐specific animal source food and multiple micronutrient powder interventions to improve the consumption of micronutrients by infants and young children in Tanzania, Kenya, Bangladesh and Pakistan

Abstract Optimal complementary feeding practices including consumption of multiple micronutrient powders (MNP) are recommended to improve micronutrient intake by infants and young children (IYC) 6–23 months. Formative research was used to design the behaviour change strategy to improve IYC micronutrient intake for the multicountry ENRICH project in rural impoverished areas of Tanzania, Kenya, Bangladesh and Pakistan. Employing a qualitative approach with multiple methods and guided by a social ecological framework, the study was conducted in several phases: data collection in the community, household and health facilities, initial analysis and household trials (HHT). Results found limited use of animal source foods (ASF) for feeding IYC and MNP largely unavailable. Although cost constrained access to ASF, potential more affordable context‐specific ASF options were identified in each setting. Caregivers associated ASF with many positive attributes for IYC, but barriers to feeding them included lack of caregiver time and knowledge of specific preparation techniques, and limited advice from health workers. Feeding practices were identified that used time‐efficient, specific preparations for eggs and other ASF, and demonstrated good initial acceptability and feasibility during HHT. Testing MNP in HHT found good initial feasibility and acceptability and provided an understanding of the facilitators and constraints for preparing, feeding and promoting MNP. In conclusion, formative research led to the design of context‐specific ASF and MNP complementary feeding promotion strategies to improve IYC consumption of micronutrients by identifying the practices, benefits, motivations and alternative actions to overcome the barriers in each setting.


| INTRODUCTION
Consumption of nutritious foods and micronutrient supplements at the household level are essential practices for infants and young children (IYC) 6-23 months in low-and middle-income countries (LMIC).
For this age group, reaching optimum micronutrient intake requires particular attention, as deficiencies and their associated morbidity occur at the same time that essential micronutrients are needed for brain development and immune functioning-critical for healthy growth and development (Bailey, West, & Black, 2015;Black et al., 2013;Cusick & Georgieff, 2016;Cusick, Georgieff, & Rao, 2018). Globally, iron deficiency represents the most common micronutrient deficiency, and iron deficiency anaemia affects an estimated 41.7% of children <5 years old, with higher regional prevalence in Africa (59.3%) and Southeast Asia (51.4%; WHO, 2020). Other common micronutrient deficiencies in IYC include vitamin A and zinc, and frequently multiple micronutrient deficiencies coexist (Bailey et al., 2015;WHO, 2016). Micronutrient deficiencies reflect a diet of poor quality, often indicated by inadequate dietary diversity at the population level (Working Group on Infant and Young Child Feeding Indicators, 2006).
Among the food groups recommended for children during the complementary feeding period to improve dietary diversity and diet quality are animal source foods (ASF), which include flesh foods (meat, fish, chicken and organ meats), eggs and dairy products (milk, cheese and yogourt) (PAHO/WHO, 2003). ASF are particularly rich in micronutrients (e.g., iron, zinc, vitamin A, vitamin B-12, among others) with higher bioavailability than plant sources as well as containing high quality protein and essential fatty acids (Bailey et al., 2015;Dror & Allen, 2011;Kwasek, Thorne-Lyman, & Phillips, 2020;Murphy & Allen, 2003). Flesh foods contain the more bioavailable form of heme iron and additionally enhance nonheme iron absorption from plant sources when consumed together (Hurrell & Egli, 2010). Yet, on a global scale, daily consumption of flesh foods (27.6%) and eggs (16.6%) are particularly low among IYC 6-23 months, and dairy products are consumed by less than half (42.2%) (White, Bégin, Kumapley, Murray, & Krasevec, 2017).
To prevent micronutrient deficiencies and promote optimal micronutrient status in IYC 6-23 months, the World Health Organization (WHO) recommends optimal complementary feeding practices and the use of multiple micronutrient powders (MNP), starting at 6 months (PAHO/WHO, 2003;WHO, 2016). To best promote these practices, intervention programmes designed with a behaviour change strategy are advised (WHO, 2016). Formative research using qualitative methods within local contexts is advocated to provide the needed data and insights to best inform behaviour change strategies for effective complementary feeding interventions (Bentley et al., 2014;Fabrizio, van Liere, & Pelto, 2014). Similarly, formative research is called for to inform context specific programming and communication around MNP (Schauer et al., 2017).

The Enhancing Nutrition Services to Improve Maternal and Child
Health (ENRICH) initiative is a multicountry intervention project aiming to improve overall maternal, newborn and child nutrition and To design these strategies, formative research was conducted to explore and understand the context-specific factors and influences around food and complementary feeding within each setting, and the feasibility and acceptability of local possibilities for behaviour change.
Specifically, this study aimed to • understand the knowledge, perceptions, attitudes and practices around complementary feeding, specifically dietary diversity, ASF and MNP and the barriers and opportunities present within households, health services and communities; • test the feasibility and acceptability of behaviours and practices to improve consumption of micronutrient rich foods such as ASF and MNP in household trials (HHT); and • recommend context-specific practices and communication (for the intervention strategies) to improve micronutrient intakes by IYC. This paper describes the formative research including HHT related to ASF and MNP for the ENRICH project.

Key messages
• Caregivers associated ASF with positive qualities for IYC.
Cost, time and preparation barriers to feeding ASF can be overcome by promoting time-efficient, doable food preparations of context-specific, less expensive choices such as small fish in Bangladesh and Tanzania, meat and milk in Kenya and chicken or fish in Pakistan.
• Eggs were underutilized across countries yet are a more available and affordable nutritious ASF to feed IYC. Negative perceptions of eggs can be modified with contextspecific preparation methods and promotion of their positive qualities.
• HHT were an important component of the methodological process for understanding the facilitators and constraints for preparing, feeding and promoting ASF and MNP.

| Study design and data collection methods
Following desk reviews of relevant health policy and previous intervention programmes in each country, the formative research commenced using a qualitative approach within a social ecological framework (Stokols, 1996). Research occurred over 1-2 months per country between November 2017 and March 2018 and included several phases and a variety of semistructured methods to engage multiple respondents as described below. Formative research template tools were created in English for each data collection method and adapted for language and context by each country's team (Table 1). Several were based on those in the ProPAN field manual (PAHO/UNICEF, 2013).
Phase 1: Initially, a free listing exercise was conducted, whereby individual participants (caregivers and various community members) listed all foods eaten in the area by families and by IYC in particular.
From these results, a 'key foods list' was created with the micronutrient rich foods mentioned, including three groups of ASF: flesh foods, eggs and milk/milk products. The market survey followed, whereby field researchers visited several local markets and shops to observe and talk with vendors about the availability, cost and seasonality of these foods. 'Food cards' picturing each of the key food were created for use in Phase 2.
Phase 2, Part A: Household research started with a qualitative dietary recall of all food eaten by the child in the previous 24 h (including approximate amounts), as well as the frequency of consuming selected foods in the past week. A semistructured, recorded interview with the caregiver followed and focused on perceptions of healthy growth, aspirations for their child, feeding practices, knowledge and perceptions of MNP and interactions with health workers. In other households, the food attributes exercise was conducted whereby caregivers pile-sorted the food cards into three groups: foods given to the child already, foods they would consider giving and foods they would not give. The field researcher then explored the specific reasons for placing each food in a group, the age to begin feeding it and its preparation. Home observations focused on meal preparation, interactions between caregivers and IYC, foods consumed, breastfeeding and the home context, using a semistructured guide.
Finally, a household food stock tool was applied to identify the foods purchased or home-produced during the past week. interviews with health workers took place at the health facility or in the community and focused on their perceptions of nutrition issues in the area, nutrition counselling, barriers and opportunities to promote optimal IYC feeding and MNP delivery.
Phase 3: After completing data collection, an initial data analysis was undertaken to identify potential priority feeding practices and foods to be addressed within each country. Phase 4: Following the methodology of ProPAN, the HHT were implemented to determine caregiver and child acceptability, feasibility and ability to carry out the feeding practices recommended from Phase 3 (PAHO/UNICEF, 2013). Three visits to selected homes in the ENRICH project sites occurred over a short period of time (two in Kenya); a qualitative 24-hour dietary recall for the child was conducted at the first and last visits. The general structure for HHT visits included an initial visit where the recommended priority practice(s) to test was selected, either assigned by the researcher or negotiated with the caregiver, followed by demonstration and discussion of the practice(s), explaining benefits and using key messages. On the second visit, the caregiver was asked to share her initial experience with the recommended practice(s) and discuss any modifications made and barriers or facilitators encountered. Encouragement to continue the practice(s) was provided. On the final visit the caregiver was interviewed about her experience testing the recommended practice(s).

| Setting and sample
The formative research community sites within each country were purposefully selected to maximize sample diversity and represent the important geographical and cultural characteristics of the ENRICH project areas, which included the following: • Tanzania: Shinyanga and Singida regions • Kenya: Elgeyo Marakwet County • Bangladesh: Thakurgaon District • Pakistan: Sukkur District, Sindh Province The various respondents and informants were purposefully selected to ensure a diverse sample for each method (e.g., different aged IYC). Sample sizes depended on country and variability within the data (Table 1). Potential participants were identified and introduced via local ENRICH staff or community health workers familiar with the communities.

| Data collection
The formative research team members underwent training and pretested the data collection tools. Standard quality assurance procedures were implemented for data collection and analysis including those for obtaining informed consent, maintaining data confidentiality and secure storage of all data. Debriefing sessions were held during the field work.

| Analysis
The recorded and transcribed interviews and observation reports were coded following the interview or observation guide topics, and adding emergent codes as needed. A thematic analysis followed using matrices to help organize, reduce, compare and synthesize all of the qualitative data (Miles, Huberman & Saldaña, 2020;PAHO/UNICEF, 2013). Triangulation across methods and informants occurred through discussion of the principal findings, aided by additional matrices to further consolidation. For this paper, we examined data and results across countries, comparing and contrasting the findings.

| RESULTS
Several foods were identified for each of the micronutrient rich food groups on the key foods list in each country for subsequent exploration. For the group of ASF, this included a number of local flesh foods, eggs, insects and milk/milk products (Table 2).

| ASF availability, cost and purchase
Market surveys demonstrated that a number of animal products were available in each country.
Most were unaffected by season, with a few exceptions (e.g., insects). Cost predominated as the main barrier to obtaining ASF; however, many families still managed to purchase or obtain fish, Exploration of egg attributes demonstrated a more nuanced and context-specific understanding compared to other ASF (Table 3). Caregivers perceived eggs to carry both positive and negative attributes for IYC, informed by a mix of nutritional properties and cultural influences.
For example, in Kenya, concerns included digestive issues for children and uncertainty around excessive protein if too much egg was consumed, or too often; yet 'rich in protein' was also considered a positive attribute of eggs. In Tanzania, food insecurity constraints manifested through some caregivers' mention that children fed eggs too often may come to expect them and behave badly when unavailable.
The demands placed on caregivers in all settings-shouldering numerous responsibilities in addition to feeding and caring for the child-emerged as another potential constraint to preparing and feeding ASF (and MNP). They voiced a preference for feeding the child foods prepared for the family instead of cooking separately. Yet they offered specific examples of preparing these family foods with additional attention and care for children such as washing spice off foods, mashing foods to soften them, and separating bones from fish.
Although ASF might be available in the family meal, its prioritization for IYC was not ensured. For example, the fish in a typical curry in Pakistan was spread thinly, and the child received little actual fish. 3.3 | ASF in the diet of IYC

| Responsive feeding
Responsive feeding during mealtime, a feeding style defined by age and developmentally appropriate interactions with IYC by caregivers, was explored because of its importance in ensuring adequate consumption of micronutrient rich foods and MNP (Bentley et al., 2014).
During the meals observed, mothers in Bangladesh were involved and helped their children to eat when time permitted; at other times this was delegated to siblings or others. However, encouragement to eat more to meet the child's age appropriate quantity was not observed.
In Tanzania

| Household trials for ASF
The ASF practices tested included several options in Tanzania (small fish, egg and milk) and Kenya (meat and egg), whereas Pakistan selected eggs to test. In Bangladesh, ASF were part of the practice tested to increase meal diversity for children. In all countries, results showed that the recommended practices were feasible and largely well accepted by mothers and children as shown in Table 4. Mothers remembered the recommended practice, almost all reported implementing the practice at least some of the time, and all intended to continue it. Cost was a limiting factor in some cases, whereby mothers reduced the frequency of giving ASF, and/or opted to give a less expensive ASF option (often eggs).  (Table 5).

T A B L E 4 Household trials with ASF by country
Country and age of IYC ASF practice tested (sample size)

Selected findings: feasibility and acceptability
Tanzania 7-16 months Feed eggs, sardines or milk daily (n = 13) • Mothers appreciated having three options. • Most IYC liked ASF except in a couple cases with specific ASF. • Mothers linked ASF to multiple nutrients for children's health.
Kenya 6-23 months Add an ASF to your child's meal • egg mixed in porridge, or fried/boiled egg • shredded meat (n = 11) • Preparing shredded meat for children was acceptable and allayed concerns about choking. • Eggs viewed a valuable alternative to meat because of lower cost. • ASF linked to child's good health, strength, growth, development. • Easy to practice (eggs).

Bangladesh 7-23 months
Provide the baby with at least 4 types of foods (including ASF such as egg, fish or chicken) (n = 6) • Caregivers found IYC liked and consumed more food when fed a diverse diet; important to help the child consume sufficient quantity at mealtime. • Mothers linked diet diversity to health and growth.
Pakistan 12-23 months Give egg to the child 3 times a week (mixed with potato) (n = 3) • Child accepted, 'tasty', 'easy to eat'. • Satisfied child 'did not feel hungry, asked to breastfeed less'. • Egg mixed with potato neutralized the 'hot' aspect of egg, becoming a good food for children. • Easy to cook.

| DISCUSSION
This formative research study identified the context-specific ASF and MNP feeding practices and communications to design a behaviour change strategy for the ENRICH project promoting increased consumption of micronutrients for IYC in rural impoverished areas of Tanzania, Kenya, Bangladesh and Pakistan. The use of a social ecological framework provided structure to investigate the barriers and opportunities affecting complementary feeding within the community, household and health services. Others have used a similar approach to capture the range of influences and determinants around complementary feeding and ASF in particular (Armar-Klemesu et al., 2018;Pachón et al., 2007;Pelto, Armar-Klemesu, Siekmann, & Schofield, 2013;Thorne-Lyman et al., 2017).
As expected, cost was a prime constraint in all settings. Nonetheless, available, lower cost ASF within the reach of many families (e.g., eggs and fish) were identified, and caregivers fed them when possible, if not daily, during the HHT and expressed their desire to continue this practice. Other studies concur with the constraint of cost but have also identified context-specific ASF to promote (Armar-Klemesu et al., 2018;Rasheed et al., 2011;Robert, Creed-Kanashiro, Villasante, Narro, & Penny, 2017;Sanghvi, Jimerson, Hajeebhoy, Zewale, & Nguyen, 2013;Thorne-Lyman et al., 2017). Results of the high regard for ASF among caregivers and their acceptability for young IYC from 6 to 7 months across settings importantly supports their promotion from the start of complementary feeding, a time period when IYC diets often lack ASF White et al., 2017). Recognizing the challenge for caregivers to adequately prepare and incorporate these foods into the child's meal lent to recommending specific, doable practices, a facilitator for behaviour change echoed by other studies. For example, in western Kenya, Ahoya, Kavle, Straubinger, and Gathi (2019) describe cooking demonstration interventions that included teaching preparation techniques such as mashing, dicing and shredding foods to encourage improved dietary diversity and ASF consumption. In Peru, food grinders were tested as a low technology solution to prepare meat and fish for IYC and allay caregiver concerns about food texture (Creed-Kanashiro, Wasser, Bartolini, Goya, & Bentley, 2018).
Feeding eggs to IYC was limited across settings and concurs with other reports of the widespread underutilization of eggs for IYC, particularly in African and Asian countries, despite their general availability and excellent nutritional profile (Iannotti, Lutter, Bunn, & Stewart, 2014;Lutter et al., 2018;White et al., 2017). Iannotti et al. (2014 and Lutter et al. (2018) draw attention to the varying cultural beliefs around eggs that may limit feeding them to IYC, especially younger IYC, but suggest that these barriers are not insurmountable, and our findings agree. By understanding the cultural and context-specific beliefs and perceptions around eggs in each country, identifying how to incorporate them into IYC diets and effectively communicate this was possible, feasible and acceptable.
Similarly, in Ecuador, formative research led to developing a successful social marketing campaign based on local culture to promote eggs (Gallegos-Riofrío et al., 2018). In India, research found that scrambling eggs overcame the negative cultural perceptions of feeding a 'hot' food to IYC (Bentley et al., 2014). In Burkina Faso, the cultural belief that children who consume eggs will become a thief exposed during formative research was targeted by using consistent, focused communication with motivational elements that spoke to local caregivers and promoted feeding eggs (Nordhagen & Klemm, 2018 Child's illness • MNP not given in one case because the child was ill (Bangladesh).
Side effects • One mother was concerned about diarrhoea, another that MNP would lead to obesity (Tanzania). One mother noted the child's stool turned black (Kenya). Forget: • Possibility of forgetting to give MNP because of busy schedules (Kenya, Bangladesh).

Storage:
• MNP sachet loss (rats consumed sachets in one Tanzanian home).
Fish were another ASF option identified to promote in three of four countries. Kwasek et al. (2020) report growing interest among LMIC in promoting fish (wild and farmed) because of its rich source of several essential micronutrients as well as high quality protein and essential fatty acids (Tilami & Sampels, 2018). Similar to our study, Thorne-Lyman et al. (2017) identified community availability, cost and household factors-including cultural beliefs and perceptions about fish for IYC, preparation time and safety as influential determinants for feeding fish to IYC in Bangladesh. Qualitative work in Kenya emphasized the importance of context whereby differing cultural perceptions and acceptance around feeding small fish to IYC were found in two regions (Hotz et al., 2015). The strengths of this study include the extent of comparable data collected in four diverse settings, guided by a social ecological framework, to thoroughly explore multiple perspectives and determinants around IYC micronutrient consumption in each context. Additionally, including HHT as part of the methodological process provided further insights on the selected practices and confidence to promote them.
The limitations of this research are recognized. Although the samples were purposefully selected to represent the diversity in these communities, project areas were extensive, and the small samples may not have captured the full range of perspectives. In Pakistan, access to health facility informants proved challenging and mealtime observations in homes were not accepted, and in Bangladesh, samples were small for caregiver interviews and home observations. HHT protocols varied by country, included small samples, and were of short duration. Although positive outcomes resulted for ASF and MNP consumption during the HHT, ongoing monitoring, support and largescale project evaluation is required to determine their sustainability.

| CONCLUSION
ASF and MNP are key sources of micronutrients for IYC, yet consumption remains limited in many LMIC settings. Formative research including HHT led to the design of context-specific ASF and MNP complementary feeding promotion strategies to improve IYC consumption of micronutrients by identifying the practices, benefits, motivations and alternative actions to overcome the barriers in rural areas of Tanzania, Kenya, Bangladesh and Pakistan. This study contributes to the literature on formative research using qualitative methods to design behaviour change interventions for improving complementary feeding and micronutrient intake in LMIC.