Promotions of breastmilk substitutes, commercial complementary foods and commercial snack products commonly fed to young children are frequently found in points‐of‐sale in Bandung City, Indonesia

Abstract Few studies have documented the marketing of commercial foods and beverages for infants and young children in West Java, Indonesia. To assess the prevalence of promotions at points‐of‐sale for commercially produced products commonly fed to young children in Bandung City, 43 small and large stores were visited in 2017. Promotions for breastmilk substitutes (BMS), commercially produced complementary foods (CPCF), and select types of commercial snack products were photographed and information recorded on promotion characteristics. There were 402 and 206 promotions observed with BMS and CPCF products, respectively. Sixteen promotions with BMS products for infants under 12 months were found in 42.9% of stores selling BMS, violating national regulations. Almost all BMS promotions (98.3%) included BMS products for ages 1 year and above (“growing‐up milks”). Of all BMS products available for sale, half of all infant/follow‐up formula and 77.2% of growing‐up milks were promoted. CPCF were found in 97.7% of stores, and 81.0% of these stores had promotions; 70.5% of all available CPCF products were promoted. Of the 2,451 promotions observed for commercial snack products, 17.3% used promotional techniques targeting young children or caregivers. Joint‐promotions were common, with BMS and CPCF marketed in combination with commercial snack products; 49.0% of BMS promotions were joint BMS‐snack promotions, and 80.0% or more of infant/follow‐up formula promotions included a commercial snack. Revising and enforcing infant food and beverage marketing regulations to ensure consistency with global standards are necessary to protect and promote optimal infant and young child feeding in Indonesia.


| INTRODUCTION
The World Health Organization (WHO) and United Nations International Children's Emergency Fund (UNICEF) recommendations for optimal infant and young child feeding include exclusive breastfeeding for the first 6 months of life, at which point safe and appropriate complementary foods should be gradually introduced to a child's diet alongside continued breastfeeding up to 2 years and beyond (World Health Organization & United Nations International Children's Emergency Fund [WHO & UNICEF], 2003). Research throughout the world has documented the health and nutritional benefits of breastfeeding  and consumption of nutrientrich and appropriate complementary foods during a critical period of growth and development (Shrimpton et al., 2001;WHO & UNICEF, 2003).
Inappropriate promotion of food products can negatively influence feeding practices and diets of infants and young children (Piwoz & Huffman, 2015;Rollins et al., 2016;WHO, 2017), and WHO has released guidelines to regulate the marketing activities of manufacturers. The International Code of Marketing of Breast-milk Substitutes CPCF must not be promoted or otherwise represented as appropriate for feeding children under 6 months and must not discourage breastfeeding; cross promotion of BMS and CPCF products is prohibited. Indonesian policy only restricts mass media advertisement of food products for infants under 12 months (MOH, 2012), with no regulation on the promotion of CPCF products at retail locations, including cross promotion with BMS, which has been documented in Indonesia (Durako & Lo, 2016 • Almost all BMS promotions included a growing-up milk and 49.0% included a joint-promotion with commercial snack products.
• Nearly one-fifth of promotions for five categories of commercial snack products were targeted to children or caregivers.
• Enforcement of national regulations is necessary to prevent promotions that can negatively impact child feeding. Additional regulations should be considered to restrict promotion of BMS for older children, jointpromotion of BMS with commercial snacks, inappropriate promotion of commercial complementary foods and child-targeted advertising of unhealthy commercial snack products. and beverages in regulations controlling inappropriate promotions of foods for young children is increasingly important as consumption becomes more prevalent (Pries, Filteau, & Ferguson, 2019) and habits and taste preferences established in childhood have been linked to long-term eating preferences (Beauchamp & Mennella, 2009;Ventura & Mennella, 2011). Indonesia lacks such regulation, in spite of the growing availability of commercially produced snack foods (Baker & Friel, 2016;Shrimpton & Rokx, 2013). An U.S. Department of Agriculture report noted $7.1 billion in sales of packaged commercial snacks in Indonesia in 2016 and that growth of the food processing industry was due in part to "aggressive promotional activities" (U.S. Department of Agriculture Foreign Agricultural Service, 2018).
Violations of the Code and national law at points-of-sale are reported in Indonesia (Durako & Lo, 2016;Hidayana, Februhartanty, & Parady, 2017;International Baby Food Action Network, 2014); however, there is little evidence on the prevalence of in-store promotions for CPCF or commercially produced snack products, their jointpromotion with BMS and whether commercial snack promotions in Indonesia target caregivers or young children. This study was conducted to document the commercial marketing that caregivers are exposed to at retail locations in Bandung City. The capital of West Java, Indonesia's most populous province, Bandung City has high under-five stunting (32.2%; MOH, 2013) and only two-thirds of children 6-23 months of age achieve a minimum acceptable diet according to WHO guidelines WHO, 2010).

| METHODS
This cross-sectional study documented availability and promotional practices in retail locations for BMS and CPCF and assessed promotions for a subset of commercial snack products commonly fed to children under 3 years of age in Bandung City.

| Sampling of points-of-sale
Sampling of points-of-sale for this assessment was based on the all carrying at least one BMS or CPCF product. Large stores were purposively sampled in consultation with local officials and nongovernmental organizations working on child health and included four grocery stores/supermarkets, four hypermarkets and two baby stores.
Seven of the locations were national or local chain stores. The 10 locations were chosen for their wide variety and volume of products that would be representative of availability in Bandung City.
For this study, small stores could include corner stores (warung/ kiosks), neighbourhood cooperative grocery stores (koperasi), minimarts and pharmacies (apotiks). Small stores were sampled for their proximity to public sector health facilities offering child health services (WHO & UNICEF, 2017). The 33 facilities were identified in coordination with the Bandung City Health Office in preparation for a survey with mothers of young children . Using Google Maps and Google Street View, four small stores in closest proximity to each health facility were listed. During data collection, stores were visited in order of proximity and the first found to sell a BMS or CPCF product was included for that facility. If a store not identified through Google Maps and Street View was found in closer proximity and met study criteria, it was used instead. In total, 19 warung/kios, 12 minimarts, 1 koperasi and 1 apotik were surveyed; 11 minimarts and the 1 apotik were national chains.

| Product definitions
This study focused on three types of commercial products: BMS, CPCF and select categories of commercially produced snacks commonly fed to young children. BMS products were defined as any formula, milk or milk-like product, in either liquid or powdered form, marketed for feeding infants and young children under 3 years of age (WHO, 2016). BMS products were sub-categorized as: 1. infant formula for infants 0-5 months of age; 2. follow-up formula for 6-11 months; and 3. growing-up milks for 12-35 months.
CPCF were foods marketed as suitable for feeding young children if they met at least one of the following criteria: (1) recommended for introduction at an age of less than 3 years; (2) labelled with the words "baby", "toddler," "young child," or synonym; or (3) in any other way were presented as being suitable for children under the age of 3 years (WHO & UNICEF, 2017). CPCF were sub-categorized into: 1. infant cereal, including instant bubur (rice porridge); 2. grain snacks/finger foods, including products such as biscuits/ cookies, puffs or rusks; 3. pureed foods and infant meals, which may include cereal, pasta, meat, poultry, fish, eggs, fruits and/or vegetables; 4. infant pudding (instant milk/gelatine pudding); and 5. other CPCF products, including tea, juice, olive oil or abon (finely shredded meat).

| Data collection
Data were collected in 43 stores during May 2017. Grocery stores/supermarkets and hypermarkets were surveyed on the weekend as this was hypothesized to be the most likely time for promotions with company representatives. The baby stores and all small stores were surveyed on week-days as previous scouting indicated that promotions appeared similar on weekends and week-days.
To guide data collection on product availability, comprehensive master lists of all BMS and CPCF products for sale in Bandung City were generated through a review of the Badan Pengawasan Obat dan Makanan (National Food and Beverage Registry), online research and informal visits to seven stores. The final master lists included product brand, flavour, product description, age of use and manufacturer.
Different flavours of a product were listed as different products. Products with varying package sizes (single-serving verses multi-serving, and different sizes of multi-serving packages) and different types of packaging (e.g. box, sachet) were listed as one product. Each product was assigned a unique identification code for data collection and analysis. Master lists for the commercial snack products were not generated because the large number of products on the market made this logistically unfeasible.
In each store, data on product availability were captured first. All areas inside a store were thoroughly surveyed (e.g. baby food section, milk section, baby supplies section and discount section) to identify all products for sale. Enumerators systematically matched each BMS and CPCF product found throughout the store to a paper copy of the master lists, checking off each individual product identified for sale. Any products not found on the master lists were added to the list for all subsequent points-of-sale to be visited.
Once all information on product availability was recorded, enumerators captured information on each promotion observed throughout the store. A promotion was defined as an individual occurrence of promotional activity in a store for one or more BMS, CPCF or commercial snack product, such as a shelf display, price discount or an informational brochure . The types of promotions assessed were: 1. price-related, such as coupons, discounts or buy-one-get-one; 2. displays, including brand shelves/counters/tables, special shop windows, posters/banners or shelf tags/talkers/wobblers; 3. information materials, like leaflets, pamphlets/brochures or catalogues; 4. free gifts to customers, such as toys, baby items or plastic foodstorage containers; 5. product samples; 6. company representatives in store; 7. store banners/signs with store name/logo and product logo/brand; and 8. any other promotions, such as contests, store bonus-points or holiday baskets.
If a promotion included two distinct promotion types, each type was counted as a unique promotion. For example, a product placed in a special display booth with an offer of a free gift was counted as one display promotion and one free gift promotion.
For each promotion, enumerators recorded the type of promotion and type of product promoted (BMS, CPCF or snack). If BMS or CPCF were promoted, the unique product codes from the master lists were recorded for each product observed in the promotion. Promotions for BMS and CPCF without distinct product sub-categories (i.e. a broad promotion for a brand/manufacturer, without indication of unique products that could be matched to the master list) were captured as "sub-category cannot be identified." For each snack product promoted, the snack sub-category (e.g. sweet biscuit), brand and manufacturer were recorded. Promotion data were collected on electronic tablets using the Android application Open Data Kit (ODK), and

| Review of caregiver-or child-targeted commercial snack promotions
After data collection was complete, photographs of the commercial snack promotions were reviewed for caregiver-or child-targeting based on seven criteria adapted from previous research (Chacon, Letona, & Barnoya, 2013;Gantz, Schwartz, Angelini, & Rideout, 2007;McGinnis, Gootman, & Kraak, 2006;Mehta et al., 2012). Criteria for child-targeted promotions included having (1) a gift for a child (e.g. toy, book and bottle) or (2) cartoons, licensed characters, images of children, sports images and/or animals on the promotion itself. Images appearing on the label of a package shown in the promotion were excluded. Criteria for caregiver-targeted promotions included having (1) the word "baby," "toddler," "young child" or synonym; (2) the words "mom," "dad" or synonym; (3) nutrient content claims (e.g. "contains iron" or "with DHA"), nutrient function claims (e.g. "promotes growth" or "improves intelligence") or health claims (e.g. "enhances immunity" or "reduces indigestion"); (4) emotional cues (e.g. "happy" or "exciting") or (5) statements directed towards caregivers about children ("we understand you only want the best for your child").
Two analysts independently reviewed all snack promotions and their corresponding photographs in accordance with protocol used by Pereira et al. (2016) and Sweet et al. (2016). Promotions were graded "Yes" or "No" for each of the seven criteria. The two sets of grades were compared for agreement by a third analyst; 2.7% disagreement was found and a third analyst made the final determination.

| Data analysis
Data on product availability were entered into Excel daily. Promotions data and photographs were reviewed by a study coordinator and then uploaded to ONA nightly. The grading of snack promotions was entered into Excel, and the check for agreement was run using Stata version 14 (StataCorp, College Park, TX, USA). Data were cleaned in Excel and all analyses were run in Stata.
The number of unique BMS and CPCF products observed across all stores was summed to calculate overall product availability. Availability was also disaggregated by sub-category of BMS and CPCF, manufacturer and location of production (domestic or imported). The number of stores selling at least one type of product or one product sub-category was calculated to report availability by store.
The number of stores with at least one promotion observed for each product type was determined. Percentage of stores with promotions was calculated by dividing the number of stores with promotions by the number of stores selling the product. Data on the availability of commercial snack products were not collected due to the overwhelming number of items available, so the percentage of stores selling snack products is reported out of all 43 stores in the study. Promotion data by store are also presented by small or large store size.
The number and percentage of available BMS and CPCF products that were found promoted across all stores were calculated by dividing the number of products promoted by the number of products found available for sale. If a product was documented in at least one promotion, it was counted as promoted. Promotion by availability was also calculated for each BMS and CPCF sub-category.
The number and percentage of all promotions found across the 43 stores were calculated by type of product and sub-category. If a promotion had at least one product type/sub-category, it was included in the count of promotions. Promotions including multiple categories of products were classified as joint-promotions, and the different combinations of product types were examined. The total numbers of BMS, CPCF and snack promotions were also disaggregated by type of promotion (e.g. price or display). Snack promotions were considered child-targeted if they included at least one of the two criteria, and caregiver-targeted if they included at least one of the five criteria. A total of 220 unique CPCF products were found. Infant cereal was the most common sub-category of CPCF found (44.1%, n = 97), followed by snacks/finger foods (32.7%, n = 72), purees and infant meals (17.3%, n = 38), puddings (3.2%, n = 7) and other CPCF (2.7%, n = 6), which included thee olive oils, two teas and one abon. were domestically produced, and most snacks/finger food products (69.4%, n = 50), purees and infant meals (92.1%, n = 35) and other CPCF products were imported (83.3%, n = 5).

| Promotions for breastmilk substitutes, commercially produced complementary foods and commercially produced snack products
BMS and CPCF products were promoted in the majority of the stores they were sold in (82.1% and 81.0%, respectively). Promotions were more prevalent for BMS and CPCF in large stores (90.0% and 100%, respectively) than small stores (77.8% and 75.0%). Almost half of stores selling BMS (42.9%, n = 12) had promotions for infant and/or follow-up formula. Nearly all stores in the study (95.3%) promoted the study's subset of commercial snack products commonly fed to children. Across all stores, a total of 402 promotions included at least one BMS product, 206 included at least one CPCF product and 2,451 included at least one commercially produced snack product (Table 1).
Nearly all BMS promotions included a growing-up milk and 16 promotions included infant formula and/or follow-up formula (4.0% of all BMS promotions). Infant cereal was the most commonly promoted sub-category of CPCF. All sub-categories of snacks were similarly promoted except for malt-beverages/non-dairy milks.
Promotions were found to promote one category of product or multiple categories of products together, such as BMS and CPCF ( Table 2).
Examples of joint-promotions included displays or posters showing different products together, company representatives promoting a range of products or buy-one get-one half-price discounts. BMS products were more frequently included in joint-promotions (197) than CPCF Although overall the most common type of CPCF promotion was price promotion, the type of promotion varied by sub-category of CPCF. Price promotions continued to be common for promotions with snacks/finger foods (68.1%, n = 32), infant cereal (57.0%, n = 87) and puddings (42.9%, n = 6). Gift promotions were highly prevalent for purees and infant meals (57.1%, n = 4) and puddings (42.9%, n = 6).
Price promotion was also the most frequently observed promotion method for all commercial snack product sub-categories, followed by display promotions.

FIGURE 1
Percentage of available breastmilk substitute and commercially produced complementary food products that were found promoted in stores. Number of products found available for sale reported as (n)

| Child-and caregiver-targeting of commercial snack promotions
In the review of 2,451 promotions with commercial snack products, 17.3% (n = 425) met at least one of the criterion for child-or caregiver-targeting. In total, 214 (8.7%) were child-targeted, 119 (4.9%) were caregiver-targeted and 92 (3.8%) were both child-and caregiver-targeted. Figure 3 shows the number and percentage of all snack promotions meeting each criterion. Promotions of candies, sweet biscuits and savoury snacks were targeted more to children (24.8%, 11.1% and 7.2%, respectively) than caregivers (6.1%, 5.4% and 3.1%). Sweetened milks and malted beverages/non-dairy milks were similarly targeted to children and caregivers: sweetened milks, 17.1% child-targeted and 22.3% caregiver-targeted, and malted beverages/non-dairy milks, 22.9% child-targeted and 18.1% caregivertargeted.

| DISCUSSION
Promotions for BMS, CPCF and commercial snacks commonly fed to young children were highly prevalent in this study of 43 stores in Bandung City, Indonesia. Over 80% of stores carrying BMS and CPCF products had promotions for these products, and commercial snack promotions were nearly universal. Nearly two-thirds of all BMS products and three-quarters of CPCF products available for sale were being promoted. The majority of BMS promotions were for growingup milks for children older than 1 year; however, we found promotions for infant and follow-up formula, which violates national regulations.
Manufacturers frequently promoted BMS jointly with commercial snack products, mainly sweetened milks. Nearly 2,500 promotions were observed for five categories of commercial snack products, with almost one-fifth targeting children and/or their caregivers. Possible for a promotion to include more than one product category or sub-category. b Sub-category of product could not be determined in the promotion.    (Sankar et al., 2015). Although the WHO recommends breastfeeding for up to 2 years or beyond, most legislation against the promotion of BMS, including Indonesia's, applies only to products through the first year of life (WHO, UNICEF, & IBFAN, 2016). Several countries with strict regulations against the promotion of BMS have seen increased rates of breastfeeding (Brady, 2012;Bragg, Hardoby, Pandit, Raji, & Ogedegbe, 2017;Champeny et al., 2016;Cyrillo, Sarti, Farina, & Mazzon, 2009). In countries like Indonesia where promotion for and use of growing-up milks is rising, there is a pressing need to expand restrictions against the promotion of BMS for children up to three years of age.
The pervasive promotions for growing-up milks found by this study is also troubling because previous research has shown that Other type of promotion 6.7 (1) 7.7 (1) 4.3 (17) 0 (0) a Data presented as percentage (n). Due to rounding, percentages may not add up to 100%. b Possible for a promotion to include more than one subcategory of breastmilk substitute. c Sub-category of product could not be determined in the promotion.
mothers and caregivers often cannot differentiate between the different stages of BMS products (Berry, Jones, & Iverson, 2010;Cattaneo et al., 2015) and caregivers confused by similar packaging could inappropriately use growing-up milks to feed younger infants. Pereira et al. (2016) documented BMS manufacturers using the same brand attributes across their range of products in four study sites. The majority of follow-up formulas and growing-up milks manufactured by infant formula companies displayed at least one example of cross promotion with one or more of the company's infant formula products: two-thirds or more contained similar colour schemes, designs and mascots. Furthermore, growing-up milk advertisements function as indirect advertising for infant and follow-up formula (Berry et al., 2010;Cattaneo et al., 2015). In the absence of regulation prohibiting cross promotion through similar packaging and labelling, the widespread retail promotion of growing-up milks in Bandung City could serve to indirectly promote a manufacturer's entire range of BMS, circumventing restrictions on BMS promotion and undermining breastfeeding.
CPCF products were found in nearly every store surveyed, indicating the ubiquity of commercially available complementary foods. Over 200 promotions were observed for 70.5% of the CPCF available for sale, which is not necessarily negative as it has long been recognized that suitable CPCF can play a role in maintaining appropriate nutritional intake for young children, especially when foods are fortified (Brown & Lutter, 2000;Dewey, 2009 foods for children Smith et al., 2015;Sweet et al., 2016), and needs to be regulated.
This study is one of the first to assess targeting of commercial snack products to children under 3 years and their caregivers in Indonesia, and the prevalence of targeting found is concerning. Promotions of commercially produced snack foods and sugar sweetened beverages have been shown to influence purchase and consumption behaviours (Boyland et al., 2016;Sonntag, Schneider, Mdege, Ali, & Schmidt, 2015). In our study, promotions for candy/chocolate/jellies and malted beverages/non-dairy milks, which are often high in added sugar, were found to most frequently use techniques that targeted children. The most prevalent caregiver-targeted promotion was for sweetened milks (22.3%), which is also troubling as a recent multicountry study on the perception of beverage healthiness found that Indonesian participants perceived flavoured, sweetened milk to be healthier than other sugar-sweetened beverages (Thomson et al., 2017). With prevalent consumption of unhealthy commercial snack products by young children in Indonesia , and rising rates of child overweight/obesity (MOH, 2013;Rachmi, Li, & Baur, 2017), promotions for these products could have a significant impact on the health and well-being of children in Indonesia. The Indonesian government should consider passing and enforcing food industry regulations related to caregiver-and child-targeted promotion of commercial snack foods to reduce the impact of unhealthy snacking among children.
This study has several limitations. Stores sampled for this study were not representative of all vendors in the city; however, WHO's recommended NetCode protocol was followed and 44.2% of the stores (n = 19) were local or national chain stores, so promotions observed were likely to be found throughout the city and country.
For commercial snack products, only a subset of categories was examined, which does not represent the totality of snack promotions or totality of promotions for commercial foods high in fat, sugar or salt in these stores. Additionally, we did not assess purchase or consumption of products so it is not known if promotion in this research led to greater consumption of these products. This assessment was carried out with assumptions about store characteristics and retail behaviour, which could have introduced bias: different store types were visited at different times/days of the week, which could have led to differentials in promotion observation, and the store sampling methodology differed between small and large store types.

| CONCLUSION
This research conducted in large and small stores across Bandung City, West Java, Indonesia, has revealed numerous promotions for BMS, CPCF and commercially produced snack products, with several BMS promotions illegal under Indonesia law. A high prevalence of childand caregiver-targeted snack promotions was found, along with frequent joint-promotion of BMS, CPCF and commercial snacks. Overall, these promotions may be contributing to decreased breastmilk intake and/or increased consumption of trans-fats, sugar and salt among young children, ultimately leading to long-term health consequences.