Peer Groups to Improve Infant Feeding Practices and Child Growth in Post-emergency Settlements in Uganda

Brief introduction:

Low-and middle-income countries (LMICs) are "home" to most of the world's refugees, the majority of whom are women and children. Refugee children are vulnerable to undernutrition and poor health, and humanitarian agencies have highlighted child undernutrition as a key area for intervention during refugee situations. Complementary feeding of infants in refugee settlements remains inadequate; however, there is limited evidence on how to address these nutritional challenges in refugee settlements. In 2020, I conducted a year-long community-based randomized control trial among post-emergency settlements in the West Nile region in Uganda to examine whether an integrated nutrition education intervention delivered through Care Groups would improve infant feeding practices and growth in the post-emergency settlements in Uganda. The findings in this study may be used to design nutrition-sensitive programs and also inform policies targeting complementary feeding practices within post-emergency settlements in Uganda.

Study area:

The Adjumani district had been randomly selected among six districts and one city hosting refugees in the West Nile region. Initially, three settlements in the Adjumani district were randomly selected and assigned to the three arms of the study. Ayilo-I settlement was assigned as a Moms-only treatment arm, Pagirinya settlement was the Moms \& Dads treatment arm while Nyumanzi settlement was the Control arm. During the recruitment of the participants, some fathers in the settlement (Pagirinya) assigned to the Moms \& Dads treatment could not commit to participating in a year-long study because they had to move away from the settlement for jobs back to Juba, the capital of Sudan or nearby towns with more employment opportunities. Because of the limited number of participants in the Moms \& Dads arm, a second randomization was done excluding the already selected settlements. Ayilo II was then selected as the other settlement to identify respondents for the Moms \& Dads treatment arm. These settlements were at least six kilometers apart to reduce the possibility of spillover effects of the intervention. The village health team (VHTs) and health center midwife assistants supported the identification of pregnant mothers to be included in the study.

Sample size:

Prompt introduction of solid and semi-solid foods to infants was used as the primary outcome to determine the sample size because of its reliability as an indicator of infant and young child feeding practices. A sample size of 317 mothers was desired for strong power, based on calculations using GPower 3.1 software, with a type I error of 0.05, a power of 0.90, and an effect size of 0.2 to detect differences in the proportions of infants introduced to complementary foods at 6-8 months among the 3 study arms. A 23 percent loss during follow-up was estimated; thus, 390 women (15-49 years of age) in their third trimester of pregnancy were willing to participate. These women were enrolled before baseline data collection. Husbands were eligible to participate with their wives in the Moms \& Dads treatment arm. A total of 321 mother-infant dyads completed the study.

Purposive sampling was used to identify pregnant women in their 3rd-trimester to be included in the study. VHTs and a midwife assistant from each settlement were contacted to support the respondent identification process due to their community roles in promoting health among expectant mothers. Identified pregnant women were crossmatched with an integrated maternity register from the health facility. Verification of pregnancy term was done with the help of the midwife assistant, VHT, and researcher by validating records in the antenatal card or maternity records book. Each of the chosen settlements formed the household sampling frame for one arm of the study; that is, each settlement had either: 1) Moms-only, 2) Moms \& Dads combined, or 3) Control arm. The participants in the study were followed up from recruitment in January 2020 to the end of the study in December 2020.

Study Intervention:

A peer-led integrated nutrition education training program using the Care Group model developed by the USAID TOPS program comprised the intervention. We hypothesized that participants in the Care Group intervention would have better peer support and would have improved infant complementary feeding and growth compared to the participants in the control arm. For the intervention arms, topics trained in the peer support groups included 1) group dynamics and social support, 2) optimal breastfeeding and complementary feeding practices, 3) adequate basic hygiene, 4) child growth and development, and 5) fathers' involvement. The training messages were adopted and modified from the community health extension workers (CHEWs) training handbook and the United Nations Children Emergency Fund (UNICEF) infant and young feeding counseling cards for community workers. Selected images were included in each module for illustration adopted from the UNICEF-IYCF counseling cards for community workers.

Intervention groups:

In this study, Moms-only and Moms \& Dads arms comprised the treatment arms also referred to as the intervention arms. The Control arm was the third study arm but did not receive the intervention. Each study arm comprised 10 groups with 10 - 20 members per group. For the study arms, volunteer leaders were selected within each group by the members. These leaders were referred to as peer group leaders. For the intervention arms, the peer group leaders were trained for five days using the same training tool that was to be used during the Care Group meetings. The peer group leaders were trained by identified village health team (VHTs) together with a nutrition and health educator from the Adjumani district. Only one refresher training for the peer group leaders was done during the entire study period even though at least two refresher trainings had been planned. The failure to conduct additional refresher training was due to measures undertaken to observe COVID-19 safety operating procedures (SoPs). Nevertheless, monthly feedback meetings were held for the peer group leaders with the VHTs to harmonize group meeting plans and address any challenges. These monthly meetings were also attended by the researcher. Each of the peer group leaders facilitated integrated nutrition training modules on a biweekly basis to the 10 to 20 group members identified in a Care Group to effect behavioral change. The group members were also encouraged to visit each other's homes as part of peer follow-up, however, caution to COVID-19 SoPs was emphasized.

Standard of care/ Control arm:

In this study, the Control arm did not receive the intervention. The Control arm comprised a standard of care that was expected to involve VHTs doing their routine work in the community supporting households concerning nutrition and health information for better maternal and child health. The GoU recommends having at least two VHTs in a village, which is between 50 and 70 households. These VHTs report directly to the nearest government health facility. Additionally, the VHTs are expected to promote government health programs and urge community members to access health-related services from government facilities.

Data Collection:

The structured questionnaire used for this study was adapted from the Demographic and Health Surveys and UNHCR standardized expanded nutrition surveys (SENS) guidelines. These two questionnaires have been widely accepted for sociodemographic, health, and nutrition data collection and related information for nationally representative samples and humanitarian populations. Additional questions to assess maternal social support were adopted from the medical outcomes study (MOS) social support scale. The final questionnaire was translated into Arabic, Dinka, and Madi languages which are widely spoken by the South Sudanese in the refugee settlements in Uganda. The questionnaires were then uploaded onto electronic tablets (Lenovo Tab P11) using the Qualtrics software (Offline version). The collected data with the electronic tablets were uploaded to Qualtrics cloud storage at the end of each data collection day. Only the principal investigator and research collaborators had access to the data as part of maintaining respondent data safety.

Four enumerators with a minimum qualification of an associate degree in nutritional science, social sciences, or related fields were recruited and trained on the questionnaire. Each enumerator was proficient in English and Arabic and one of the local languages (Madi, Nuer, or Dinka). A total of 10 households with women pregnant in their 3rd trimester were used to pretest the questionnaire in the Boroli refugee settlement. The Boroli refugee settlement where the pretesting was done was not among the settlements that were selected for the year-long study. The researcher was able to make minor revisions and adjustments mostly regarding the flow of questions in the Qualtrics software. Data collectors/enumerators were blinded from the study arm assignment. The collection of data/information from a single respondent was completed within 1 hour and 30 minutes. The data collection time points for this study were baseline (January), Midline-I (June), Midline-II (October), and Endline (December). At each data collection period, the participants were provided with a 1 kg bar of washing soap, 200 mL of vitamin A fortified cooking oil, and half a kilo each of iodized salt and sugar, all worth about 7,600 Ugx (1.5 USD) as compensation for participation in the study.

Ethical Considerations:

This study was approved by the Institutional Review Boards of the Uganda National Council of Science and Technology (SS 5038), Makerere University School of Health Science Research and Ethics Committee (SHSREC REF:2019-020), and Oklahoma State University (HS-19-2). Additional permission was acquired from the Office of the Prime Minister (OPM) Uganda (OPM/R/107). Informed consent was obtained from all respondents at recruitment for participation in the study, and before data collection during the study.