Menu
Menu
Menu
Menu
Background: Optimal infant and young child feeding practices (IYCFP) reduce childhood stunting and are associated with additional health benefits. In Tanzania, IYCFP are far from optimal where 32% of children under the age of 5 years are stunted. The purpose of this study was to examine whether behavior change communication focused on reducing child undernutrition was associated with improved IYCFP in Tanzania. Methods: A cross-sectional survey was administered to approximately 10,000 households with children under the age of 2 at baseline and endline. Bivariate analyses and logistic regression was used to examine the relationship between exposure to behavior change communication and timely initiation of breastfeeding, exclusive breastfeeding, continued breastfeeding at one year, timely complementary feeding (CF), minimum meal frequency (MMF), minimum dietary diversity (MDD), and minimum acceptable diet (MAD). Results: Mothers who heard a radio spot about IYCFP were more likely than mothers who had not heard a radio spot about IYCFP to begin complementary foods at six months. Their children were also more likely to achieve MMF, MDD, and MAD with odds ratios of 2.227 (p = 0.0061), 1.222 (p = 0.0454), 1.618 (p = <.0001), and 1.511 (p = 0.0002), respectively. Mothers who saw a TV spot about IYCFP were more likely to have greater odds of knowing when to begin complementary feeding, feeding their child a minimally diverse diet (4 food groups or more), and serving a minimum acceptable diet with odds ratios of 1.335 (p = 0.0081), 1.360 (p = 0.0003), and 1.268 (p = 0.0156), respectively. Conclusion: Exposure to behavior change communication in Tanzania was generally associated with some increased knowledge of optimal IYCFP as well as practicing IYCF behaviors. Behavior change communication planners and implementers may want to consider conducting similar campaigns as an important component of behavior change to reduce undernutrition and poor health outcomes in developing settings.
From 2015–2020, IMA World Health (IMA) implemented the Department for International Development-funded “Scaling up Growth: Addressing Stunting in Tanzania Early (in the under 5’s)” (ASTUTE). The project was designed in close collaboration with the Tanzanian Ministry of Health. ASTUTE was specifically designed to support the Tanzania’s nutrition strategy, including a mandate to build the capacity of District Nutrition Officers (DNuOs) to manage and coordinate, at the district level, the nine nutrition-relevant government sectors through Council Multisectoral Steering Committees for Nutrition (CMSCN), mirroring the coordination at the national level by the High Level Steering Committee on Nutrition (HLSCN). The National Institute for Medical Research provided ethical clearance. The behavior change communication was implemented in five regions of the Lake Zone in Tanzania (see Fig. 1). These are Geita, Kagera, Kigoma, Mwanza, and Shinyanga with a collective population of 10.2 million and over 750,000 stunted children. These regions were selected for their documented high prevalence of stunting and anemia and poor infant and child feeding. The behavior change communication focused on three major objectives: Map of intervention area. This map was provided by the ASTUTE program and is used with permission In the designated regions, an evidence-based communication campaign was implemented between June 2017 and March 2020, which included radio and TV campaigns (17.6 million reached), mobile outreach (8.4 million reached through direct messaging), and IPC in the form of home visits (6.4 million reached). The radio spots were theory-based and lasted 60 s. They were broadcast 10 times a day for a total of 70,000 times. TV spots were also 60 s and aired a total of 1, 198 times. They aired on three different national/regional stations before and during the news. A mother’s knowledge about IYCFP, as well as environmental and social influences are important determinants of her nutrition related behaviors. Social Cognitive Theory (SCT) is a well-established theoretical approach that may be utilized to address parental feeding practices and to inform behavior change communication development [14]. The constructs of SCT are cognitive influences, environmental influences, and supporting behavioral factors such as self-efficacy [15]. Many communication campaigns utilize techniques from SCT by modeling desired behaviors on television using actors that are culturally or ethnically similar to the audience [11]. When done correctly, behavior change strategies can help increase the self-efficacy of the audience, and can address inappropriate cultural or social practices that adversely affect childhood nutrition status [11, 12]. The use of SCT allows for a more thorough understanding of nutrition-related behaviors [16, 17]. The ASTUTE program utilized a cross-sectional survey that was distributed to 5,000 households before the behavior change communication was implemented and an additional 4,996 households after the behavior change communication was implemented. Inclusion criteria included having a child under two years (0–23 months) of age and living in the regions where the campaign took place. Respondents who did not meet these criteria were excluded from the study. Survey questions were directed to the female caregiver of the youngest child in the household, and if available, to the male head of household. Consent was received before the survey was administered, and participants understood that their participation was voluntary, they could stop at any time, and the potential risks and benefits associated with their participation. The survey was developed in English, and then translated to Swahili. It was piloted and edited, and ultimately included 169 questions which aimed to measure participants’ exposure to the communication campaign and other outcomes. Data were collected by a field team consisting of 50 enumerators and 10 supervisors. All field team members received a two-week training prior to participant recruitment and data collection. Their goal was to recruit 5,000 households during three rounds of surveys. Survey participants were selected using a stratified, multi-stage random sample design. During the baseline round, 243 villages were included and participants were randomly sampled within each village. During the following two rounds of surveys the same villages were used, but participants were again randomly selected. Interviews were conducted in the participants’ homes and lasted on average 50–60 min. Data were collected digitally using smartphones and PDAs (personal digital assistants). At baseline, 5,000 female caregivers and 1,114 male heads of household were surveyed from January to February 2017. At endline, 4,993 female caregivers and 3,084 male heads of household were surveyed from January to February 2020. While the survey used in this study was not validated as part of this work, it was largely based on previously validated instruments such as the Demographic and Health Surveys and was pilot tested in the field before baseline data were collected. Behavior change communication objectives analyzed during the study included the reach and exposure of the campaign, change of key indicators, and the association between exposure and key indicators. Authorization for this research and intervention was obtained from the Ministry of Health in Tanzania and Development Media International’s (DMI) internal IRB. Data quality was checked by 11 controllers, and if the quality of a previously completed interview could not be validated, a new interview was conducted. Additionally, raw data was checked for outliers and invalid answers. Female caregivers provided demographic and household information in the survey to understand the distribution in each sample group. Before data analysis occurred, a single binary definition was created for each variable of the campaign. Primary and secondary outcome variables were also defined for each campaign message theme. A wealth index variable was created to estimate relative household income. This variable was based on the index created by Briones [18]. It includes household access to services such as safe water and sanitation as well as ownership of goods including radio, TV, bicycle, motorcycle, automobile, mobile phone, boat, and animal-drawn cart. The score is the average of the services and goods scores. Values range between 0 and 1, with wealth increasing as the value gets closer to 1. Two female and two male media exposure variables were created. Female exposure to radio was defined as ‘yes’ if women reported recalling a radio message that discussed maternal nutrition, exclusive breastfeeding, or child nutrition after six months. Female exposure to TV was defined as ‘yes’ if women reported recalling a TV message about maternal nutrition, exclusive breastfeeding, or child nutrition after six months. TV and radio variables were also created for males using the same methodology. A variable was created to measure overall exposure to the behavior change communication. This variable, which only includes responses from mothers, assesses whether the respondent had no exposure to the behavior change communication, heard or watched any behavior change messages (media only), had any IPC-related interactions (IPC only), or both media and IPC. Seven variables measuring feeding practices were used, in alignment with World Health Organization (WHO) standards for IYCF indicators [10]. They included timely initiation of breastfeeding, exclusive breastfeeding, continued breastfeeding at one year, timely complementary feeding (CF), minimum meal frequency (MMF), minimum dietary diversity (MDD), and minimum acceptable diet (MAD). These variables are described below. Timely initiation of breastfeeding, as defined by the WHO, means beginning breastfeeding within the first hour of life [10]. Putting baby to breast immediately or within the first hour was considered timely. The question used to create this variable was “how long after birth did you first put (name) to breast?”. For exclusive breastfeeding, the primary outcome was defined as “proportion of mothers of children 0–6 months who are currently breastfeeding and report they haven’t given the child any other food/liquids.” This was assessed by asking participants whether they are breastfeeding and to select food and drinks they had fed their infant within the last 24 h. Continued breastfeeding at one year was assessed by asking mothers of infants ages 12–15 months whether they are “still breastfeeding (name).” For complementary feeding, mothers of children ages 6–8 months were asked: “Have you introduced (name) any other fluids or foods besides breast milk?” “How old was (name) when he/she was first fed something other than breast milk?”. MMF is defined by the WHO as being fed 2 times per day for breastfed infants 6–8 months, 3 times for breastfed children 9–23 months, and 4 times for non-breastfed children 6–23 months [2]. Based on these standards, participants were asked the age of their child and whether they are breastfed. They were then asked if the child ate anything when they woke up in the morning, anything between then and lunch, anything at lunch, anything between lunch and dinner, anything at dinner, and anything after dinner. This variable measured how many food groups were represented in the child’s (6–23 months) diet in the previous day. Participants were asked to select whether their child had eaten specific types of grains, legumes, dairy, flesh, eggs, and fruit/vegetables. Children who ate from four or more of these groups were coded as having MDD. If children ages 6–23 months achieved MDD and MMF, then they were coded as having a MAD. The raw data were cleaned, recoded if necessary, and analyzed using SAS version 9.4. The baseline data were first compared to the endline data to determine change for the key indicators. A chi-square test was conducted to determine whether these differences in key indicators. between baseline and endline were statistically different. A multiple logistic regression model was built to determine the relationship between exposure to the media campaign and increased IYCFP among female caregivers. The model controlled for maternal and male head of household age, education, child age, and wealth.
N/A
