Evidence of Effective SBCC Approaches to Improve Complementary Feeding Practices

Background

At the age of six months, breast milk alone is no longer sufficient to meet the nutritional requirements of a growing child (World Health Organization (WHO)/UNICEF, 1998). Timely and appropriate complementary feeding is critical to a child’s growth and development. According to the 2008 maternal and child undernutrition series in The Lancet, the adoption of optimal complementary feeding practices and other supportive strategies at scale could result in a 17% relative reduction in the prevalence of stunting at 24 months, and avert 5.5 million disability-adjusted life years (DALY) (Bhutta, 2008a). In a similar series in 2013, The Lancet reiterated the association between appropriate complementary feeding and nutritional outcomes, citing a study by Ruel and Menon (2002) that found a positive, statistically significant association between feeding practices and height-for-age in five Latin American countries. Finally, in an analysis of datasets from 14 low income countries, Marriott et al. (2012) reported that “consumption of a minimum acceptable diet with dietary diversity reduced the risk of both stunting and underweight whereas minimum meal frequency was associated with lower risk of underweight only.”

Despite the promise of timely appropriate complementary feeding, global coverage of optimal complementary feeding practices remains low. According to an analysis of 2002- 2008 data on complementary feeding practices in 46 countries, just one-third of children aged 6-23 months received the recommended dietary diversity,⁴¹ approximately half received the recommended number of feeds, and only 20% received the minimum acceptable diet (Lutter, 2012).⁴²

This chapter focuses on the effectiveness of Social and Behavior Change Communication (SBCC) approaches to improving complementary feeding practices of children 6-24 months of age. The chapter focuses on the WHO guiding principles for complementary feeding for the breastfed child (Pan-American Health Organization (PAHO)/ WHO, 2003).See Table 3.3.1.

Search Results

Thirty peer-reviewed studies met the literature review inclusion criteria.⁴⁴A complete list of the studies reviewed, including a summary of the SBCC approach(es) utilized, the study design, practices measured, and levels of significance is provided in Table 3.3.5.

Twenty-five studies reported a statistically significant change in at least one of the recommended complementary feeding practices. Each of those articles is described in greater detail below. Studies reporting a statistically significant change in more than one of the prioritized practices and/or utilizing more than one SBCC approach are cited and discussed multiple times. Where this occurs, we have provided a detailed description of the study in the first reference, and only discuss results in subsequent references.

The findings are organized by practices targeted, SBCC approach(es) utilized, and study design. This organization enables the reader to derive an independent judgment regarding findings from each study. In addition, implementation processes followed are discussed.

Of the 30 studies reviewed for this chapter, eleven were randomized controlled trials (RCT), nine were longitudinal studies (five with control groups and four with no control group), and ten were repeated cross-sectional studies (five with control groups and five with no control group). See Table 3.3.2.

[block:views=706fcfe999d71c53f4f78dc67a568ca1]

Overview of the Evidence, by Practice

With the release of the PAHO/ WHO’s Guiding Principles for Complementary Feeding of Breastfed Children (2003), the need to develop standardized indicators to capture the breadth of recommended complementary feeding practices was established. This recognition resulted in the development of five standardized indicators of complementary feeding: introduction of solid, semi-solid or soft foods, minimum dietary diversity, minimum meal frequency, minimum acceptable diet (MAD), and consumption of iron-rich or iron-fortified foods (WHO, 2008).

The studies reviewed in this chapter, however, revealed very little consistency in how complementary feeding practices were measured, and few employed the standardized indicators described above. The review is therefore focused on the WHO Guiding Principles rather than the “recommended” indicators. See Table 3.3.1. The guiding principle related to the safe preparation and storage of food was considered beyond the scope of this review.

Of the 30 studies reviewed, five measured the introduction of complementary foods (mean age at or the frequency of introduction of foods at specified ages); 20 explored the nutrient content of foods; five studied the use or consumption of fortified products or supplements; 16 measured meal frequency, energy density, or energy intake; three looked at how children were fed (responsively or actively), and three measured feeding practices during and after illness. See Table 3.3.2.

Timely introduction of appropriate complementary foods

Of the five studies looking at the introduction of complementary foods or particular food types, two reported at least one statistically significant result. Each of the studies measured the practice differently, and neither reported on the recommended standard indicator, namely the “proportion of infants 6-8 months of age who receive solid, semisolid or soft foods” (WHO, 2008).

De Oliveira et al. (2012) conducted an RCT measuring the mean age of introduction of complementary foods in a hospital in Brazil. The intervention included client-provider counseling in the facility and follow-up home visits. At the conclusion of the study period, the mean age at introduction of complementary foods was 153 days in the intervention group, compared with 95 days in the control group. Similarly, 22.8% of children were introduced to complementary foods at four months of age in the intervention group, compared with 41% in the control group. However, the authors found that “at 6 months, the prevalence of infants receiving complementary foods was similar for the two groups: 87.1% for the intervention group and 88.4% for the control group.”

Li et al. (2007) conducted a repeated cross-sectional study looking at the mean age of introduction of seven weaning foods in five townships in Yunan Province, China. The intervention included one-on-one counseling, food preparation demonstrations, peer education, growth monitoring from birth through 17 months, and dissemination of a maternal and child nutrition video and pamphlet. The project also supported the strengthening of integrated management of childhood illness (IMCI)-related services, and provided thiamine to women just before or after delivery. Following the intervention, the age of introduction of all food groups except cow’s milk had increased.

Dietary diversity

Of the 20 studies that looked at the nutrient content of complementary foods, 15 reported statistically significant results. Of those reporting statistically significant findings, 13 reported on the feeding of animal source foods, nine reported on the feeding of fruits and vegetables, and three reported on dietary diversity. Just one article (Sun et al., 2011) reported on the recommended standard indicator, “proportion of children 6-23 months of age who receive foods from 4 or more food groups.” (WHO, 2008)

Feeding children animal source foods

Of the 14 studies that measured the feeding or consumption of animal source foods, all but one reported statistically significant results. The outcomes measured, however, varied considerably among studies. There was little consistency among studies with respect to the meat product promoted (meat, cow’s milk, buffalo milk, eggs, positive deviance (PD) foods including animal source foods, and/ or “iron-rich foods”), the measurement of milk intake (quantity consumed and frequency of consumption), and measurement of egg consumption (mean number of times consumed and percentage of children fed three or more eggs). Furthermore, reference periods and age ranges varied considerably among studies.

Randomized controlled trials

Seven of the studies reporting statistically significant results were RCTs. Bhandari et al. (2004) looked at the consumption of milk, fruit, and vegetables in India. The intervention involved “monthly home visits for new births until aged 12 months and weighing once every 3 months for children under 2 years of age conducted by Anganwadi workers, immunization clinics run by the auxiliary nurse midwives and sick child contacts with health care providers.” Additionally, auxiliary nurse midwives discussed key messages during monthly meetings with community representatives who, in turn, shared them during monthly neighborhood meetings. Following the intervention, the percentage of children nine months of age consuming any type of milk was 91.8% in the intervention group, compared with 82.6% in the control group. This finding was statistically significant. While the difference at 18 months of age was not statistically significant, differences in mean energy intake from milk between the intervention and control group were statistically significant at both 9 month and 18 months – 837kcal vs. 607kcal at 9 months of age and 2021kcal vs. 1289kcal at 18 months of age.

Bortolini and Vitolo (2012) looked at mean intake of cow’s milk and meat among children 12 to 16 months of age in Rio Grande do Sul, Brazil. Women in the intervention group received monthly home visits during their child’s first six months of life, and follow up visits when the children reached 8, 10 and 12 months of age. Following the intervention, the proportion of women postponing the introduction of cow’s milk was higher in the intervention group (50.9%) than the control group (36.6%). Furthermore, mean intake of cow’s milk and meat in children 6-12 months of age was higher in the intervention group (588.1ml and 54.3 g, respectively) than in the control group (501.6 ml and 47.3 g, respectively).

Guldan et al. (2000) looked at the frequency of feeding children eggs, meat, and fish or meat broth in rural Sichuan, China. The intervention included counseling by community health workers (CHW) during monthly growth monitoring sessions, and visits to the homes of pregnant women and women with children under the age of one. During the home visits “nutrition educators disseminated a feeding guidebook and growth chart to each family, [provided] age-appropriate breastfeeding and complementary feeding suggestions and advice, answered questions and weighed each infant, marking the infant’s weight on the growth chart.” Following the intervention, the percentage of children ages 4-6 months and 7-9 months that were fed eggs was 37% and 66% respectively in the intervention group, compared with 16% and 41% respectively in the control group. The percentage of children ages 7-9 months and 10-12 months that were fed fish or meat broth on a daily basis was 9% and 13% respectively in the intervention group, compared with 1% and 3% respectively in the control group, while the proportion of children ages 7-9 months and 10-12 months that were fed meat was 19% and 23% respectively in the intervention group, compared with 8% and 9% in the control group.

Penny et al. (2005) looked at the consumption of egg, chicken liver, or fish in Trujillo, Peru. The intervention aimed to “enhance the quality of nutrition counseling through training and provision of simple, standardized, age-appropriate messages to be used at all points of contact with young children [or their caregivers] in the facility.” Data were collected on a cohort of children from intervention and control facilities from birth to age 18 months. Following the intervention, a higher proportion of women in the intervention group reported feeding egg, chicken, liver, or fish across the six measurement points (at 6, 8, 9, 12, 15, and 18 months of age) than women in the control group; however, results were only significant at six and eight months of age. The practice of feeding animal foods improved in the immediate period following counseling; but diminished over time.

Roy et al. (2007) looked at the feeding of eggs, the addition of oil to children’s food, the feeding of energy and protein rich local complementary foods rich in micronutrients,⁴⁵ and the feeding of children from a separate pot in Bangladesh. The intervention included the delivery of nutrition education sessions to small groups of women through trained community health workers. The groups met weekly for the first three months and biweekly for the second three months. Formative research guided development of messages, and the sessions included food preparation demonstrations. In addition to the mothers groups, men and older family members were targeted through monthly community/social mobilization efforts. Following the intervention, the percentage of children receiving three or more eggs increased from 11.3% to 29.8% in the intervention group and from 10% to 11.9% in the control group. Furthermore, the percentage of caregivers reporting feeding an animal-based complementary food increased from 30.4% to 88.5% in the intervention group and declined from 31% to 24.5% in the control group.

Vazir et al. (2013) looked at consumption of iron-rich foods among children 9-15 months of age in 60 villages in India. The intervention, implemented over a period of 12 months, included regular home visits by trained village women (twice a month or four times a month depending on the age of the infant). The complementary feeding group (CFG), received eleven nutrition education messages on sustained breastfeeding and complementary feeding, while the responsive complementary feeding and play group (RCF&PG), received these same messages as well as eight messages on responsive feeding and eight messages on developmental stimulation using five simple toys. The control group received routine services provided under India’s Integrated Child Development Services program.

At nine months of age, 30.7% of children in the CFG and 25.6% of children in the RCF&PG consumed goat or chicken liver during the previous week compared with 2.3% of children in the control group. In addition, 22.3% of children in the CFG and 11.3% of children in the RCF&PG consumed goat meat during the previous week compared with 4.5% of children in the control group. At 15 months of age, 43.5% of children in the CFG group and 44% in the RCF&PG group consumed liver compared with 13.1% of children in the control group, while 37.5% of children in the CFG group and 45.2% of children in the RCF&PG group had consumed poultry in the previous week compared with 18.9% in the control group. The findings were statistically significant. The authors found no statistical difference with respect to intake of eggs.

Finally, Zaman et al. (2008) looked at the feeding of eggs, meat, and liver, the addition of ghee or butter to children’s foods, and feeding of a fortified complementary food in Lahore, Pakistan. The intervention was introduced as part of an IMCI program, and included counseling of sick women and children seeking care at a local health center. The counseling was provided by trained health workers equipped with counseling cards that had been adapted to the local culture and language. Eight to 14 days after recruitment, the proportion of women offering their children eggs was 16 percentage points higher among children in the intervention compared with those in the control group. Six months after recruitment, the difference was 21 percentage points. The proportion of women offering children chicken, beef, or mutton 8-14 days after recruitment was approximately 18 percentage points higher in the intervention group than the control group. These findings were statistically significant. The authors found no statistically significant differences in the feeding of chicken, beef or mutton six months after recruitment, or in the feeding of liver during any period of follow up.

Longitudinal studies

Pachón et al. (2002) conducted a longitudinal randomized study looking at the consumption of locally available PD foods in rural Vietnam.⁴⁶ The intervention was based on the PD/Hearth approach.⁴⁷ The authors identified several locally available highly nutritious foods, which were then promoted in intensive nutrition rehabilitation sessions for mothers of malnourished children. Women attended the sessions for two weeks each month for up to nine months. During the intensive intervention period (months 2-6) children under the age of three in the intervention group were fed PD foods an average of 4.1 times per day compared with an average of 3.6 times per day among children in the control group.

Palwala et al. (2009) conducted a longitudinal study without controls looking at mean intake of cow’s milk, eggs, meat, legumes, fruits, and vegetables in Mumbai, India. Trained fieldworkers worked with groups of 8-10 mothers using what the authors describe as “innovative modules and demonstrations.” The workers then followed up weekly using “a simple checklist … to assess impact, identify practices not adopted, and provide further inputs.” By the end of the study period, mean intake of milk among children 6-36 months of age had increased from 134 mL to 202 mL, while mean intake of egg had increased from 20 g to 29 g, and mean intake of meat (mutton, chicken, or fish) had increased from 4 g to 9 g.

Repeated cross-sectional studies

Aboud et al. (2008) looked at mean egg, fish, and cow’s milk consumption among children in Sripur, Bangladesh. Both intervention and control groups received 12 education sessions on child development, including five on complementary feeding. The intervention group received an additional six sessions on responsive feeding. Although both the intervention and control groups participated in sessions on complementary feeding, at the five month follow up children in the intervention group consumed a mean of 0.28 eggs during the previous 24 hours, while children in the control group consumed just 0.09 eggs in the previous 24 hours. The difference was statistically significant. Differences in the consumption of fish and cow’s milk were not significantly different.

Mackintosh et al. (2002) conducted a repeated crosssectional study with controls looking at consumption of specific foods,⁴⁸ hygiene, and health practices in Tan Hoah, Vietnam. The intervention included growth monitoring and promotion (GMP) for all children under three years of age, and a PD/Hearth nutrition rehabilitation program for malnourished children that involved education and demonstration with their caretakers. At the conclusion of the study period, 93.3% of children in the intervention group had consumed shrimp in the previous week compared with 76% of children in the control group, and 97.8% of children in the intervention group had consumed eggs in the past week, compared with 72% in the control group. These findings were statistically significant. Similarly, the mean frequency of shrimp and egg consumption in the intervention group (3.2 and 4.0, respectively) was significantly greater than consumption in the control group (1.0 and 1.5, respectively). Differences in the consumption of meat and fresh fish were not statistically significant.

Sun et al. (2011) measured the percentage of children fed iron-rich foods (defined here as meat, fish, poultry, liver and iron supplements) before and after an eight month behavior change communication (BCC) intervention in Shan’xi, China. The program, which was primarily aimed at increasing children’s consumption of a multiple micronutrient powder, included formative research and development of messaging and materials. The BCC materials included infant and young child feeding (IYCF) handbooks for parents, booklets for health workers, and television spots. Caregivers “were educated about [IYCF] (including the use of YYB) during their children’s regular village doctor visits. Brochures were also distributed by pediatricians to families with infants and young children during sick and well child visits at maternal and child health care hospitals.” Data from the endline survey revealed that the percentage of children fed iron-rich foods had increased from19.2% to 56.8%; however, the lack of baseline values or a control group makes it difficult to attribute these differences to the intervention.

Cross-sectional studies

Ghosh et al. (2002) looked at the consumption of animal milk, the number of food groups consumed, and feeding frequency among children in a cross-sectional study conducted in 13 villages in India. The intervention included monthly home visits by trained workers for a period of four to six months. Comparisons were made between infants included in the baseline study who reached 12 months of age before the intervention began (control group) and those who were less than six months old at the start of the intervention and continued to participate until the child reached 12 months of age (intervention group). By the end of the study period, 53.2% of children between six and 10 months of age in the intervention group had been fed animal milk in the previous 24 hours compared with 31.9% of children in the control group.

Feeding children fruits and vegetables

Of the ten studies that measured fruit and vegetable intake, nine reported statistically significant results. The indicators measured varied considerably between studies, making comparisons across studies difficult.

Randomized controlled trials

Bhandari et al. (2004) conducted an RCT looking at vegetable and fruit consumption among children in India. At the end of the study period, the percentage of children nine months of age consuming any type of vegetables was 6.7% in the intervention group, compared with 2% in the control group. This difference was statistically significant. The authors found no statistically significant differences between the intervention group and control group with respect to fruit consumption, or children’s consumption of vegetables at 18 months of age.

Guldan et al. (2000) looked at the consumption of fruits and vegetables in rural Sichuan, China. Following a peer counseling intervention, the percentage of children in the intervention group consuming fruit exceeded the percentage among controls in all age groups (26% vs. 10% at 4-6 months of age, 68% vs. 42% at 7-9 months of age, and 79% vs. 52% at 10-12 months of age). These findings were statistically significant. Vegetable consumption among children in the intervention group was significantly greater than the control group in children 7-9 months of age (79% vs. 65%), but not in children 4-6 or 10-12 months of age.

Finally, Vazir et al. (2013) looked at the consumption of bananas and spinach by children in India. The study included two different intervention groups: a CFG and an RCF&PG. At nine months of age, 60.9% of children in the CFG and 59.4% of children in the RCF&PG consumed bananas during the previous week, compared with 38.6% of children in the control group, while 20.1% of children in the CFG and 14.4% of children in the RCF&PG consumed spinach during the previous week, compared with 5.1% of children in the control group. Both findings were statistically significant. At 15 months of age, 79.3% of children in the CFG group and 78.3% in the RCF&PG group consumed bananas, compared with 61.9% of children in the control group, while 42.4% of children in the CFG group and 45.2% of children in the RCF&PG group had consumed spinach in the previous week, compared with 29.5% in the control group.

Longitudinal studies

Pachón et al. (2002) children under three in the intervention group were fed PD foods,⁴⁹ including vegetables, green vegetables, sweet potatoes, bananas, oranges, papayas, and tangerines, an average of 4.1 times per day compared with only 3.6 times per day among children in the control group. This finding was statistically significant.

Palwala et al. (2009) measured children’s mean consumption of fruits and the proportion of children consuming vegetables in Mumbai, India. By the end of the study period, mean intake of fruit in children between six and 36 months of age had increased from 41g to 82g, the percentage of children consuming green leafy vegetables had increased from 16.7% to 87.6%, and the percentage of children consuming carrots had increased from 8.3% to 52.1%. The results were statistically significant.

Repeated cross-sectional studies

Aboud et al. (2008) conducted a repeated cross-sectional study looking at mean fruit and vegetable intake among children in Sripur, Bangladesh. Five months after the responsive feeding intervention, the mean number of times children consumed fruit and vegetables during the previous day was .6 and 1.37 (respectively) in the intervention group, compared with .32 and .8 (respectively) in the control group.

Mackintosh et al. (2002) looked at consumption of greens, sweet potatoes, and fruit in Thanh Hoa, Vietnam. At the conclusion of a GMP and PD/Hearth intervention, 100% of children in the intervention group consumed fruit in the previous week, compared with 88% in the control group. Similarly, the mean frequency of fruit consumption in the previous week was 5.4 in the intervention group, compared with 1.5 in the control group. These findings were statistically significant. Differences in the consumption of greens and sweet potatoes were not statistically significant, although the authors did report an increase in the mean frequency of sweet potato consumption (6.7 times/week in the intervention group vs. 3.2 times/week in the control group).

Parvanta et al. (2007) looked at child feeding practices, particularly associated with iron-rich foods, in northwest Bangladesh. Intervention group 1 (IG1) included homebased counseling (once in a three month period) by health educators to pregnant women and mothers of children 6–24 months old. Intervention group 2 (IG2) included discussion of role modeling stories facilitated by peer group leaders during gardening group or credit group meetings. Participants were asked to share these stories with related women who fell into the eligibility categories mentioned above. Following the intervention, 52% of women in IG2 reported feeding children 6-24 months of age the recommended portion of green leafy vegetables, compared with 33% of women in IG1, and 17% of women in the control group.

Cross-sectional studies

Finally, Kilaru et al. (2005) looked at the consumption of bananas and dietary diversity in rural Karnataka, India. The intervention included monthly counseling (referred to as “education” in the article) for caregivers of infants from birth through a child’s second birthday. The nutrition education was provided in the home by trained field research staff (high school or college educated fieldworkers who live in close proximity to the study villages). “After administering the monthly questionnaire, the field staff would discuss reported feeding practices with the primary caregiver. The total time for the questionnaire and counseling was about 1½ hours. Field staff would review dietary information from the questionnaires and talk about ways of improving the quality and quantity of the diet through questions and probes asked in a friendly manner. The field staff would often provide personal examples of how to overcome challenges in feeding young children, including the kind of communication skills needed to coax young children to eat.” Following the intervention, 33% of caregivers in the intervention group reported feeding children 7-11 months of age bananas, compared with 4% in the control group.

Feeding children a minimum number of food groups (dietary diversity)

Two of six studies that measured dietary diversity reported statistically significant results. Sun et al. (2011) conducted a repeated cross-sectional study looking at dietary diversity in Shan’xi, China. Following a provider counseling and media intervention, the percentage of women reporting that they fed their six to 24 month old children more than four food groups⁵⁰ had increased from 57.5% to 74.1% among breastfed children, and from 27.6% to 53.9% among nonbreastfed children. Furthermore, the percentage of children fed a minimal acceptable diet had increased from 42.1% to 73.5% among breastfed children and from 25.0% to 45.0% among non-breastfed children. There was no control group included in the study.

Kilaru et al. (2005) conducted a cross-sectional analysis looking at dietary diversity in Karnakaka, India. Following a home-based counseling intervention, the percentage of children 11 months of age consuming at least five different food groups was 42% in the intervention group, compared with 19% in the control group.

Micronutrient supplementation or fortification

Although we identified many studies reporting on the consumption of micronutrient supplements or fortified products, only five presented results in a way that enabled an examination of the distinct effect of specific SBCC approaches on uptake. Two of these reported statistically significant results.

Bonvecchio et al. (2007) conducted an RCT looking at the impact of a behavior change intervention designed to improve the utilization of a multi-micronutrient nutritional supplement (papilla) among children 6-23 months of age in Mexico. Papilla was provided as part of the Oportunidades Program free of charge, but conditional upon mothers’ participation in monthly education sessions on health and nutrition in all study communities. In intervention communities, these education sessions were complemented by home visits in which volunteers conveyed four key messages related to the preparation and use of papilla, and demonstrated its preparation. Health care service providers conveyed similar messaging during consultations, and videos on these topics were screened in health facility waiting rooms. Posters were mounted in multiple sites, including health centers, stores, community centers, and churches, while megaphones mounted on trucks broadcast messages and announcements. At follow up, the authors reported a net increase of 43.5 percentage points in proper preparation of papilla in intervention communities compared with control communities, a net increase of 64.4 percentage points in daily provision of papilla, and a net increase of 61.5 percentage points in those providing papilla between breakfast and dinner.

Dongre et al. (2011) conducted a repeated cross-sectional study with controls focused on increasing coverage of iron supplementation among children under the age of three in 23 villages in India. The intervention included nutritional education sessions conducted during the Bal Suraksha Diwas (child health day) celebration and during home visits, free distribution of iron syrup, and monthly village-level community/ social mobilization meetings. Furthermore, “a photograph of locally available iron-rich foods and citrus fruits was shown to mothers to increase their awareness of sources of iron. […]” After three months, the percentage of children under the age of three who took iron supplementation increased from 8.8% to 41.6%. It is important to note that this intervention distributed the iron syrup free of charge.

Increased meal frequency and/or energy density

Sixteen studies measured meal frequency and energy density and 12 of them reported statistically significant results. Four broad categories of practices were reported: seven studies reported on frequency of feeding, four reported on the addition of fats or oils to children’s foods, two reported on enhancing the energy density of children’s foods (i.e. decreasing water content), and three reported on increasing quantity fed to children at each meal.

Increasing frequency of feeding children

Seven of eleven studies that looked at the frequency or number of meals fed to children over a specified period of time reported statistically significant results.

Randomized controlled trials

Bhandari et al. (2004) looked at meal frequency among children at nine and 18 months of age in India. At the conclusion of the facility and home based counseling intervention, children nine months of age in the intervention group consumed a mean of 4.4 meals in the previous 24 hours, compared with a mean of 3.9 meals in the control group, while children 18 months of age in the intervention group consumed a mean of 5.9 meals in the previous 24 hours, compared with a mean of 5.4 among children in the control group. Both differences were statistically significant. Mean energy intake in the previous 24 hours was higher among children in the intervention group than in the control group (1556 kilojoule (kJ) of energy vs. 1025 kJ among 9 month olds, and 3807 kJ vs. 2577 kJ among 18 month olds).

Roy et al. (2007) looked at children fed three or more times per day in Bangladesh. At the conclusion of the group education and small media intervention, 83.8% of children in the intervention group were fed three or more times per day compared with 19.4% of children in the control group. These differences remained statistically significant at follow up six months later (88.5% vs. 24.5% respectively).

Longitudinal studies

Pachón et al. (2002) looked at meal frequency among children under the age of three in rural Vietnam. After two to six months of a PD/Hearth group education intervention, children in the intervention group were fed a mean of 4.6 times per day compared with a mean of 4.2 times per day among children in the control group. After 12 months of the intervention, children in the intervention group consumed a mean of 4.9 meals per day compared with a mean of 4.6 meals per day among children in the control group. Furthermore, during the intensive intervention period (months two to six) children in the intervention group were more likely to meet energy intake requirements (kcal/kg) than those in the control group (49.0% vs. 35.1%). This difference was statistically significant. At the 12 month follow-up period, however, the difference in the percentage of children meeting the recommended intake was not statistically significant.

Repeated cross-sectional studies

Guyon et al. (2009) looked at the impact of interventions designed to increase food intake during pregnancy and lactation and feeding frequency in children in two highland provinces in Madagascar. The intervention included training of health workers, interpersonal communication, community/ social mobilization, and the use of mass media. At the end of the study period, “the proportion of breastfed children 6-23 months of age fed the recommended number of times increased from 87% to 92%. There was no increase in the rate of feeding the recommended times per day among children 6-8 months of age, but there was a significant increase among children 9 to 23 months of age (from 85% to 93%).”

Kimmons et al. (2004) reported on several studies assessing the feasibility of different approaches for improving complementary feeding practices in rural Bangladesh. Each study was conducted for seven days among approximately 30 women. One tested the feasibility and effectiveness of educational messages to increase the frequency of meals fed to children. Women were asked to feed at least three meals per day, and more if possible. They were advised that snack foods (amounts less than 10 g) were not to be considered meals, and they were encouraged to continue to give snacks as usual. By the end of the seven day trial, the number of meals provided to children in the previous 24 hours had increased from a mean of 2.2 to a mean of 4.1. The second study focused on increasing the quantity of food offered to, and consumed by, infants at each meal. This trial did not have an effect on meal frequency, but did improve the quantity of food consumed by children (see below in discussion of quantity fed to children at each meal). In the third small study, women were taught to decrease viscosity of children’s food (and therefore increase food or energy consumption) by “changing the recipe or adding amylase, increasing energy density by adding sugar or oil, or changing the food’s organoleptic characteristics.”

Mackintosh et al. (2002) conducted a repeated cross-sectional study looking at feeding frequency among children in Tan Hoa, Vietnam. Following a PD/Hearth intervention, children in the intervention group consumed a mean of 2.9 full meals per day compared with a mean of 2.2 meals per day among children in the control group, and children in the intervention group consumed a mean of 2.5 snacks per day compared with a mean of 1.6 snacks per day in the control group.

Cross-sectional studies

Kilaru et al. (2005) looked at the frequency of feeding in two age groups in India. Following a provider counseling intervention, the percentage of children 7 to 11 months of age that were fed at least four times in addition to breast milk in the previous 24 hours was 78% in the intervention group and 51% in the control group.

Adding fats or oils to children’s foods

Five studies looked at the addition of oil, butter, or other fat to children’s diets to increase energy density, and all but one reported statistically significant results. Bhandari et al. (2004) conducted an RCT looking at the addition of oil or butter to infant foods in India. Following an intervention that included home visits and counseling at the facility level, the percentage of children nine months of age consuming foods with added oil or butter was 8.6% in the intervention group compared with 0.5% in the control group, and the percentage of children 18 months of age consuming foods with added oil or butter was 24.1% in the intervention group compared with 5.8% in the control group. The researchers found a corresponding change in energy intake from oil (kJ per 24 hours) among children nine months and 18 months of age.

In a second RCT, Mackintosh et al. (2002) conducted a repeated cross-sectional study in Tan Hoa, Vietnam. Among other foods, they looked at the consumption of oil/fat by children between 7-9 months of age. Following a PD/ Hearth intervention, children in the intervention group were fed oil/fat 6.6 times per week on average compared with 2.8 times among children in the control group.

Roy et al. (2007) looked at the addition of oil to children’s food in Bangladesh. Following the intervention, the percentage of children at six months of age that received food with added oil was 69.8% in the intervention group compared with 20.9% in the control group, and the percentage of children at 12 months of age that received food with added oil was 61.3% in the intervention group compared with 21.5% in the control group.

Finally, Vazir et al. (2013) looked at the addition of fat to complementary foods and the energy intake of children in India. The study included two different intervention groups: a CFG and an RCF&PG. At nine months of age, 51.4% of children in the CFG and 44.4% of children in the RCF&PG consumed added fat during the previous week compared with 24.4% of children in the control group. At 15 months of age, 42.4% of children in the CFG and 52.5% of children in the RCF&PG consumed added fat during the previous week compared with 29.5% of children in the control group. Similar differences in median energy intake following intervention were also reported. At nine months of age, median energy intake among children in the CFG group was 348 kcal and 331kcal in the RCF&PG group, compared with 209 kcal in the control group. At 15 months of age, median energy intake among children in both intervention groups was 569 kcal compared with 460 kcal among children in the control group.

Enhancing the consistency of children’s foods

Four studies evaluated the effect of an SBCC intervention on energy density of complementary food by enhancing the consistency of foods. Two reported statistically significant results. Zaman et al. (2008) conducted an RCT looking at the feeding of kitchuri, a thick complementary food in Lahore, Pakistan. At 8 to 14 days after recruitment, 61.6% of women in the intervention group had offered their children kitchuri compared with 45% of women in the control group, and at 180 days after recruitment, 65.9% of women had offered their children kitchuri compared with 44.3% in the control group. Findings 8 to 14 days after recruitment were not statistically significant, while findings from 180 days after recruitment were statistically significant.

Penny et al. (2005) conducted an RCT looking at the feeding of thick consistency foods among six age groups in Trujillo, Peru. A significantly higher percentage of caregivers in the intervention group compared with the control group fed food of a thick consistency to their children at measurements taken at six months of age (31% vs. 20%), at nine months of age (35% vs. 17%), and at 12 months of age (42% vs. 26%). Differences at the ages of 8, 15, and 18 months were not statistically significant.

Increasing the quantity fed to children at each meal

While several studies measured energy intake among children, only three reported statistically significant results specifically relating to increasing the quantity of food fed to children at each meal. Two studies reported on the use of a separate feeding pot, and one article reported on an increase in the quantity of food offered and time allocated for feeding children.

Roy et al. (2005) conducted an RCT looking at the use of a separate feeding pot for malnourished children 6 to 24 months of age in rural Chandpur, Bangladesh. The intervention took place within the Bangladesh Integrated Nutrition Project (BINP), and included two intervention groups of mothers. In the first group, women received intensive nutrition education (INE group) twice each week for three months. In the second group, women received the same nutrition education as well as supplementary feeding (INE+SF group). The control group received standard BINP services of twice-monthly nutrition education. After three months, 99% of women in the INE group and 100% of women in the INE+SF group reported the use of a separate pot for children, compared with 47% in the control group. At the end of six months of observation, 100% of women in both intervention groups reported use of separate pots for children, compared with 48% in the control group.

Roy et al. (2007) also looked at the use of a separate feeding pot for children 6 to 9 months of age in four upazilas (administrative units within districts) in four geographic regions of Bangladesh. Following the intervention, the authors reported an increase of 14.3% in the proportion of children served from their own pots.

Finally, Kimmons et al. (2004) conducted a longitudinal study looking at the quantity of food and duration of feeding in Bangladesh. Following an intervention in which field workers encouraged women to increase the quantity of food fed to their children and allocate more time for meals, the researchers observed no statistically significant difference in the amount of food offered. However, they did observe a statistically significant difference in the quantity of food consumed between day one of the study (40 g) and day 7 (64 g), and in the duration of the meal – from six to 13 minutes.

Responsive and active feeding

Three studies looked at responsive or active feeding and two (both RCTs) reported statistically significant results. Bhandari et al. (2004) looked at several active feeding practices in Indian children at nine and 18 months of age. The intervention included provider counseling, peer counseling, and group education to mothers. At nine months of age, 34.8% of women in the intervention group reported active encouragement, compared with 7.7% in the control group. At 18 months of age, 89.7% of women in the intervention group reported active encouragement, compared with 49.0% in the control group. Differences between the intervention group and control group held across a series of specific practices, including feeding with love and affection (41% vs. 11%), trying repeatedly if the child did not eat (15.4% vs. 1%), woman feeding her child herself (34.9% vs. 26.4%), woman making the child sit in her lap (18.2% vs. 6.6%), and feeding the child with other family members (12.4% vs. 2.5%).

Zhang et al. (2013) looked at active feeding and feeding during illness among infants in rural Lasihui, China. Children were enrolled at two to four months of age and were assessed at 6, 9, 12, 15 and 18 months of age. The intervention included group training sessions conducted by trained township doctors for mothers, grandmothers, and fathers, and home visits by trained providers every three months “with the purpose of identifying possible feeding problems and providing individual counselling.” The project disseminated child feeding booklets, and conducted demonstrations of food preparation using locally available, affordable, acceptable and nutrient-dense foods. After 18 months of the intervention, 88.0% of women in the intervention group encouraged their children to eat when they refused food, compared with 53.9% in the control group.

Feeding during and after illness

Three studies measured practices related to complementary feeding of sick or malnourished children, but only one reported statistically significant results. Zhang et al. (2013) conducted an RCT looking at feeding of breast milk and easy-to-digest foods to sick children in rural Laishui, China. Following the intervention, 88.8% of women reported feeding breast milk and easy-to-digest foods to their children with diarrhea, compared with 65% of women in the control group, while 67.3% of women in the intervention group reported preparing easy-to-digest food separately for their children with diarrhea, compared with 32.8% in the control group.

Summary of the Evidence

SBCC approaches utilized

For the purposes of analysis, we organized SBCC activities into three primary categories or approaches: interpersonal communication, media, and community/social mobilization. These approaches and the specific activities associated with each are described in further detail in the introductory chapter.

The studies reviewed included a broad range of approaches to promote optimal complementary feeding practices. Most included more than one specific SBCC activity and many utilized more than one approach.

Of the 25 studies reporting statistically significant findings, 24 utilized interpersonal communication (IPC) approaches, 12 utilized some form of media, and five employed community/ social mobilization. Group education in communities was the most commonly used IPC approach; small media was the most commonly used media approach; and issue groups were the most commonly used forms of community/ social mobilization. See Table 3.3.3.

The breadth of evidence regarding the effectiveness of SBCC in changing complementary feeding practices varies by type approach. Below, we discuss the evidence associated with each of the primary SBCC approaches utilized to promoted complementary feeding practices. Table Table 3.3.4. provides a consolidated summary of these findings.

Interpersonal communication

We identified no previous reviews looking at the effectiveness of IPC on uptake of optimal complementary feeding practices, but did identify 24 primary research studies reporting statistically significant outcomes associated with interventions that included IPC. One study reported a positive effect with respect to the timely introduction of complementary foods, 15 reported a positive effect with respect to dietary diversity, two documented a positive effect with respect to supplementation or fortification, 12 reported a positive effect with respect to meal frequency or density, two reported a positive effect with respect to responsive or active feeding, and one reported a positive effect with respect to feeding during or after illness. Twelve of these studies measured the effectiveness of IPC independent of other SBCC approaches, including four RCTs. Indicators and study design varied considerably between studies, making consolidation of findings difficult. In the only study comparing the effectiveness of different IPC approaches in changing complementary feeding practices, Parvanta et al. (2007) compared individual stage-based counseling by a trained agent and education through storytelling led by a group leader. They found that “twice as many women who met with the trained agent (in contrast to those who met with the group leaders) demonstrated feeding a portion of green vegetables equal to or surpassing the recommended quantity.” Though statistical significance of this difference was not reported, it was suggested in the authors’ analysis.

Media

We identified no previous reviews documenting the effectiveness of media-related interventions on complementary feeding outcomes, but did identify 13 primary research studies that reported positive outcomes associated with interventions including media-based SBCC approaches. One study reported a positive effect with respect to the timely introduction of complementary foods, nine studies reported a positive effect with respect to dietary diversity, one study documented a positive effect with respect to supplementation or fortification, seven reported a positive effect with respect to meal frequency or density, one study reported a positive effect with respect to responsive or active feeding, and one study reported a positive effect with respect to feeding during or after illness.

None of these studies, however, measured the effect of media interventions independent of other SBCC approaches, making it difficult to make definitive conclusions regarding the effectiveness of media related interventions on complementary feeding outcomes.

Community/social mobilization

We identified no reviews documenting the effectiveness of community/social mobilization approaches on complementary feeding outcomes, but did identify five primary research studies reporting statistically significant outcomes that included community/social mobilization approaches: One study reported a positive effect with respect to the timely introduction of complementary foods, one reported a positive effect with respect to dietary diversity, one documented a positive effect with respect to supplementation or fortification, two reported a positive effect with respect to meal frequency or density, one reported a positive effect with respect to responsive or active feeding, and one reported a positive effect with respect to feeding during or after illness.

None of these studies measured the effectiveness of community/ social mobilization independent of other SBCC approaches. It is therefore difficult to derive conclusions regarding the effectiveness of community/social mobilization with respect to uptake of complementary feeding practices.

Multiple SBCC approaches

We identified no reviews documenting the effectiveness of multiple SBCC approaches on complementary feeding outcomes. Of the 25 primary research studies reporting statistically significant results with respect to complementary feeding outcomes, one study reported a positive effect with respect to timely introduction of complementary foods, seven reported a positive effect with respect to dietary diversity, two reported a positive effect with respect to micronutrient supplementation and fortification, six reported a positive effect with respect to meal frequency and/or density, one reported a positive effect with respect to responsive or active feeding, and one reported a positive effect with respect to feeding during or after illness.

No studies compared the effect of different SBCC approaches or measured the effect of the utilization of more than one SBCC approach compared with the utilization of just one approach. One must use caution, therefore, in deriving conclusions regarding the effectiveness of interventions employing multiple approaches vis-à-vis those employing just one approach.

Implementation processes followed

None of the studies reviewed in this chapter assessed the effect of implementation processes (e.g. intensity and timing of communications; type and training of person communicating messages, or target audience) on the effectiveness of the SBCC approach on complementary feeding practices, nor did they consistently report all aspects of the implementation processes.

The intensity of communications ranged from one to 24 visits/sessions, conducted weekly or monthly for various lengths of time. However, none of the studies compared the effect of timing or frequency.

In terms of target audiences, all of the studies reviewed in this chapter targeted the mother or primary caregiver of the child. Among the 25 studies reporting statistically significant results, fourteen targeted only the mother/caregiver while eleven targeted an additional audience as well. Eight also targeted direct influencers, eight targeted local community actors, and one targeted actors of the enabling environment. None of the studies compared the effect of targeting one vs. multiple audiences or of targeting different audiences.

Conclusions

The evidence of the effect of SBCC on complementary feeding practices is quite broad, and clearly indicates that SBCC interventions can improve a wide range of complementary feeding practices. While the literature may reflect a bias to publish positive results, it also underscores the important role of SBCC approaches in improving nutrition practices – practices which have been shown to have an impact on nutritional status of women, infants, and children.

There is very little consistency in how complementary practices are defined, the normative standards followed for them, and the indicators used to measure them. This is the case, even with the existence of globally recognized indicators and measurement guidance from the WHO. An additional challenge is that behaviors are highly context-specific when it comes to complementary feeding – with different food taboos and different foods available that are appropriate for complementary feeding. If behaviors aren’t standardized, a project may improve behaviors but it will be difficult to attribute changes in outcomes to those behaviors.

Similarly, there is considerable variation in the description of SBCC interventions – the interactions or combinations with other interventions, target groups, content, messages, scale and coverage, length and intensity, as well as context.

SBCC interventions to promote complementary feeding practices are suited to iterative programming, because even if standards for behaviors aren’t met, there can be significant movement toward the standard. In a hypothetical example, the target number of families introducing leafy green vegetables at six months might not be met, but many families may have begun introducing by eight months, whereas before the intervention they were not introducing until one year. Iterative interventions can learn what supported that move in the right direction, and build on them.

Interpersonal communication is the most prevalent SBCC approach used and researched, and is consistently effective at improving complementary feeding practices. All but one of the studies in this chapter used IPC, and half of these measured the effectiveness of IPC independent of other SBCC approaches. While media and community/ social mobilization were used, they were always used with at least one other communication approach. No studies compared the effectiveness between different approaches or between one versus multiple SBCC channels or specific activities. This may be due to the difficult and cost of implementing multiple arms of a study or disaggregating the contribution of single channels or specific activities within a multi-channel intervention. Whatever the reason, it is important for SBCC practitioners and researchers to consider whether this line of research is worth the effort it will require.

All of the studies reporting statistically significant results targeted the mother of the child herself. More than half of those targeted one or more other audiences or influencers as well. No studies were identified that compared the effect of targeting one vs. multiple audiences or of targeting different audiences.

Little was mentioned in the articles reviewed regarding the timing of communications. The intensity of communications ranged from one to 24 visits or sessions, conducted weekly or monthly for various lengths of time. None of the studies compared the effect of timing or intensity of communications.

Most studies were implemented on a relatively small scale, within only a few communities and typically having fewer than 300 study participants per group.

Finally, differences in local context (including social norms, culture, and environmental factors) as well as differences in the implementation and scale of implementation affect the success of interventions. This underscores the importance of proper context assessments, formative research and/or ethnographic study prior to SBCC implementation.

Due to the lack of standardization in the way research related to SBCC is designed and described, it is challenging to make conclusions beyond the fact that projects with SBCC will result in uptake of promoted practices. While there is much to be learned from this body of literature to aid us in developing future programs, a number of questions remain. These include questions related t

• the positive (or negative) effect of using multiple SBCC approaches compared with focusing on only one;
• the positive (or negative) effect of targeting multiple audiences or influencers of the behaviors being promoted, rather than focusing on just one target population;
• the role of context, in other words, the effect of the same SBCC intervention implemented in different contexts;
• the effectiveness of different approaches (including intensity and targeting) for different behaviors;
• the cost and cost effectiveness of various SBCC approaches (particularly as it relates to scalability); and
• the effectiveness and sustainability of these approaches when implemented at scale.