Menu
Menu
Menu
Menu
Background: Large-scale social and behavioral change communication (SBCC) approaches can be beneficial to achieve improvements in knowledge, attitudes, and practices (KAP). Addressing Stunting in Tanzania Early (ASTUTE) included a significant SBCC component and targeted precursors to stunting including KAP related to maternal and child health, antenatal care, WASH, childhood development, and male involvement. METHODS: Baseline, midline, and endline surveys were conducted for a total of 14,996 female caregivers and 6726 male heads of household in the Lake Zone region of Tanzania. Regression analyses were used to estimate differences in KAP from baseline to midline and endline. Results: Women’s knowledge of handwashing and infant/child feeding practices, and attitudes related to male involvement, consistently improved from baseline to midline and baseline to endline. Women’s practices related to antenatal care, breastfeeding, and early child development improved from baseline to midline and baseline to endline. Improvements in KAP among male heads of household were varied across indicators with consistent improvement in practices related to child feeding practices from baseline to midline and baseline to endline. Conclusion: Many changes in KAP were observed from baseline to midline and baseline to endline and corresponded with SBCC programming in the region. These results provide support for the value of large SBCC interventions. Public health efforts in settings such as Tanzania may benefit from adopting these approaches.
UKaid and the Foreign, Commonwealth and Development Office (FCDO) provided funding to IMA World Health for the implementation of ASTUTE. A consistent tagline was used at the end of each theory-based radio spot, which was broadcast a total of 70,000 times. TV spots were aired before and during the evening news on national and regional stations a total of 1198 times. CHWs used a problem-based negotiated behavior change approach during in-home visits to implement IPC components of the intervention. They counseled mothers and referred children with growth faltering to health facilities for treatment and counseling. They also encouraged both mothers and male partners to engage in stimulation activities (e.g., drawing, playing, playing, naming objects, or talking with them) for their children by providing education and support. Data came from three distinct cross-sectional surveys completed by a unique sub-sample of participants during each data collection period between 2016 and 2020. Surveys were conducted in five regions of Tanzania’s Lake Zone region, namely Geita, Kagera, Kigoma, Mwanza, and Shinyanga. A stratified, multi-stage random sample design was used to select survey participants. Eligibility to participate was limited to households with a child under two years of age. Participants were randomly sampled from 243 villages that were selected from among the five participating regions. The baseline survey was carried out in 2016, prior to the launch of ASTUTE programming. A total of 5000 mothers, hereafter known as female caregivers, and 1144 corresponding fathers, hereafter known as male heads of household, were surveyed. The midline survey was conducted in 2019 and included 5000 female caregivers and 2502 male heads of households. The endline survey was conducted in 2020 after all ASTUTE programming ended and included 4996 female caregivers and 3080 male heads of household. The present study sample includes all baseline, midline, and endline participants for a total of 14,996 female caregivers and 6726 male heads of household. Participant demographics are presented in Table 1. The female caregiver of the youngest child in the home responded to questionnaire items. The male head of household was asked to respond only if available and applicable. IPSOS, a local research firm, collected all three waves of data. They comprised a field team with 50 enumerators and 10 supervisors. Twenty-five percent of records were quality-checked using revisits and phone checks. DMI’s internal IRB and Tanzania’s National Institute for Medical Research (TZ: NIMR/HQ/R.8a/Vol.IX/2344) provided Institutional Review Board (IRB) approval. Informed consent was collected before the surveys began and participants were reminded that participation was voluntary and they could stop the survey at any time. Questionnaire items were written in English, translated into Kiswahili, and then back-translated to English to ensure the original meaning was retained. The questionnaires were piloted, modified, and finalized before being administered to participants. Interviews were conducted in the participants’ homes and lasted on average 50–60 min. Baseline data were collected using hard copies and midline and endline data were recorded using smartphones and PDAs (personal digital assistants). Participants’ demographic characteristics were measured and collected. Exposure to the various components of the intervention (radio, TV, and IPC intervention in the midline and endline questionnaires) were also collected along with key MNCH indicators. Wealth. A calculated composite variable adapted from a previously validated index was used to estimate household wealth [19]. Two sub-indices comprised the index. The first sub-index represented access to services and ownership of consumer durables was the second. Items pertaining to access to services included the availability of safe drinking water sources (e.g., protected wells, a public standpipe) and safe sanitation (e.g., a flush toilet). Pit latrines were not considered safe sanitation for this study. Seven items were measured to represent consumer durables. These included ownership of a radio, TV, bicycle, motorcycle, mobile phone, boat, or animal-drawn cart. Each index was calculated by summing the total of the indicators within each index. An average of the two indices was then used to calculate an overall wealth score, with possible values ranging between 0 and 1. Higher wealth scores indicate higher socioeconomic status. Housing quality was not included in this index as the data were not available. Intervention Exposure. Only data collected at endline were used to measure exposure to the intervention. Exposure was estimated separately for each of the radio, TV, and IPC intervention components. Exposure to the radio component was coded ‘yes’ if respondents reported affirmatively to having heard the example spot(s) that concluded with the sound of a laughing baby or if they reported having heard radio messages that advised about maternal and child health and/or child development. Exposure to TV was coded ‘yes’ if respondents reported affirmatively to having seen the example image frame(s) on TV or ‘reported seeing messages on the TV that advised about maternal/child health/child development’. IPC exposure was coded ‘yes’ if respondents reported affirmatively that a (community) health worker had visited their home and advised them about maternal and child health and/or child development. Exposure to each intervention component (radio, TV, and IPC) was measured for female caregivers. IPC mostly targeted female caregivers, so male head of household respondents were only asked questions about exposure to radio and TV. Data were deidentified and shared only with study personnel to ensure confidentiality. STATA version 16 (College Station, TX, USA) was used to clean and recode variables in each of the three datasets. SAS 9.4 (Cary, NC, USA) was used to conduct analyses. Basic frequency statistics were calculated for key demographic variables. Logistic regression analysis was used to identify changes in KAP at each time point, comparing the midline and endline values to the baseline values. All models were adjusted for respondent age, education, and household wealth.
N/A
