Information Diffusion and Social Norms Are Associated with Infant and Young Child Feeding Practices in Bangladesh

"Despite the critical role of optimal infant and young child feeding (IYCF) practices in improving child health, growth, and development..., optimal practices have been adopted slowly and their adoption process is poorly understood."

Interaction among individuals within social networks can contribute to spreading information and translating messages into social norms. However, evidence on the process of shifting nutrition-related social norms - toward actual behaviour change - through the diffusion of information is limited. In Bangladesh, from 2009 to 2014, Alive & Thrive (A&T) implemented social and behaviour change communication (SBCC) interventions, including interpersonal counseling (IPC), community mobilisation (CM), mass media (MM) campaigns, and advocacy, to promote adequate infant and young child feeding (IYCF) practices and create an enabling environment for mothers and caregivers to adopt the recommended practices. This study aimed to advance understanding of how mothers' social networks, diffusion of information, and social norms link receipt of intervention messages and IYCF practices.

Intervention implementation has been described in detail elsewhere (see, e.g., Related Summaries, below). Briefly, 20 rural "upazilas" (subdistricts) with the Bangladesh Rural Advancement Committee (BRAC)'s essential healthcare programme were randomly selected to receive A&T's intensive interventions (including intensified IPC, CM, and MM) or nonintensive interventions (standard IPC with less-intensive CM and MM). BRAC frontline workers and volunteers conducted routine home visits and provided standard information on IYCF practices in both intensive and nonintensive areas. Also:

• IPC was based on multiple age-targeted IYCF-focused visits to households with pregnant women and mothers of children aged years and younger by the frontline workers and volunteers, as well as home visits by a nutrition-focused frontline worker called Pushti Kormi (PK).
• CM included awareness-building activities such as meetings and forums with husbands, religious leaders, health committee members, and other community leaders, as well as community theatre shows focused on IYCF topics.
• The MM component, which consisted of 7 television spots containing emotionally appealing mini-dramas with various messages to spread the perception of recommended practices as normative, was televised nationally 12–24 times/day, 3 times/week during prime viewing slots. In intensive areas, added screening of the same TV spots along with a cartoon film on IYCF ("Meena") was shown via community-based video shows in areas where TV reach was lower, accompanied by community dialogue and quiz shows.

The interventions were developed by incorporating ideas from behavioural (e.g., stages of behavioural change, reasoned actions, intentions and norms, interpersonal interactions, self-efficacy, and learning from role models) and community models (e.g., diffusion of innovation through social networks). A theory of change was developed based on formative research about constraints on feeding practices and how to address them.

The researchers conducted household surveys at endline in 2014 and follow-up in 2016 (n = 2,001 at endline and n = 2,400 at follow-up). They used multiple regression to test differences and changes in networks, diffusion, and norms within intervention areas. For example, to assess the extent of IYCF information diffusion, they asked mothers whether they had ever heard specific messages related to breastfeeding (such as early initiation of breastfeeding, feeding of colostrum, no prelacteal feeding, exclusive breastfeeding (EBF), and no water or other liquids up to 6 months) and complementary feeding (such as feeding mashed family foods and animal-source foods, and cooking children's food with oil). They also analysed paths from intervention exposure to IYCF practices through networks, diffusion, and norms.

In short, they found that:

• Mothers' networks of known adopters of IYCF practices were larger in intensive than in nonintensive areas in 2014 and increased in both areas over time [25-38 percentage points (pp)].
• The percentage of mothers who knew other mothers who had adopted optimal IYCF practices in the community was higher in intensive compared to nonintensive areas at endline 2014.
• The percentage of mothers who shared IYCF messages with other people was higher in intensive than in nonintensive areas in 2014 and increased significantly over time in both areas. In both areas, more family members and health workers provided IYCF information over time. Sharing of information increased 17-23 pp in intensive and 11-41 pp in nonintensive areas over time. Increased receipt and sharing of IYCF information in both areas at follow-up, particularly higher in nonintensive areas, indicated a wider diffusion of information over time.
• Perceived descriptive norms (beliefs about the prevalence of a specific behaviour, i.e., what other people do) improved 8-16 pp in intensive and 17-28 pp in nonintensive areas. Perceived injunctive norms (beliefs regarding the degree of approval of specific behaviors, i.e., what other people think) were high in both areas.
• Breastfeeding practices were associated with networks, diffusion, and norms (odds ratio (OR): 1.6-4.4 times larger comparing highest with lowest quartile).
• Minimum dietary diversity was associated with larger networks and diffusion (OR: 1.5-2.2) but not with social norms.
• The indirect effects from intervention exposure to practices show that, for minimum dietary diversity, 34%, 42%, and 43% of the total effects of IPC, MM, and CM, respectively, were explained by improved social networks, diffusion of information, and social norms. The indirect effects for IPC were 61% for early initiation of breastfeeding and 39% for EBF; the indirect effect of MM was 78% for EBF.

Thus, 2 years after termination of the initial donor's support, mothers' networks of known adopters, receipt and sharing of IYCF information, and social norms remained high in intensive areas and increased substantially in nonintensive areas. Differences in social networks, diffusion of information, and social norms, in turn, explained 34-78% of the total effects. Two years after endline, given BRAC's continued implementation of IYCF interventions as part of its routine essential health care programme services, mothers' awareness of IYCF messages was maintained in intensive areas, and BRAC frontline health workers continued to be a major source of IYCF information in those areas. Evidence of a wider diffusion of information beyond BRAC health workers was also shown through increased IYCF information received through other sources (other health workers and family members) in both areas and increased sharing of IYCF information by mothers, particularly with other women in their family and in the community. Diffusion of IYCF information in the community was corroborated by a larger network of known adopters.

As noted here, in the context of rural Bangladesh where collectivism (such as being a member of a family, extended family, or extended relations in a community group) is common, the role of social norms on breastfeeding practice is particularly important to consider. Notably, the current study only measured social norms as reported by mothers and not by other members of their social network. Additional measures of large and complex social networks could provide more comprehensive insight on how information and norms spread in the community.

In conclusion: "Diffusion of IYCF information through social networks, reinforced by positive social norms for messages promoted over time, will contribute to positive changes in IYCF practices that may be achieved and sustained through large-scale social and behavior change interventions."