EFFECTIVENESS OF A COMMUNITY-BASED PARTICIPATORY NUTRITION PROMOTION PROGRAM TO IMPROVE CHILD NUTRITIONAL STATUS IN EASTERN RURAL ETHIOPIA: A CLUSTER RANDOMIZED TRIAL

Embargo until
2019-05-01
Date
2015-04-02
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
Background: Appropriate complementary feeding and hygiene may reduce child undernutrition. To improve child nutritional status and feeding and hygiene behaviors in rural Ethiopia, a community-based participatory nutrition promotion (CPNP) program was designed, adapting positive deviance theory and involving two-week group nutrition sessions. This program was implemented to complement the existing Essential Nutrition Action (ENA) and Community-based Management of Acute Malnutrition (CMAM) programs. The effectiveness of the CPNP program in improving child growth and caregiver feeding and hygiene practices was tested in a cluster randomized, controlled trial using the existing programs alone as the control group. To better understand how the CPNP achieved expected outcomes, we evaluated the following five key intervention components, using a program impact pathway (PIP) analysis approach: (1) quality of the CPNP implementation (2) participants’ acceptance of the program, (3) participants’ recall of delivered messages (4) child feeding practices and nutritional status at the community level, and (5) the influence of CPNP on participation in the existing ENA and CMAM programs. Methods: The CPNP program was launched on August 2012 in Habro and Melka Bello districts of Oromia region of Ethiopia. The districts were divided into 12 clusters of which the CPNP was implemented in 6, while the ongoing programs of ENA and CMAM were carried out in all 12 clusters. Out of 2,064 (1,032 per group) child and mother pairs randomly selected from the locally prepared child rosters in Habro and Melka Bello districts, 1,790 children 6 to 12 months of age and their mothers, 914 and 876 pairs in the control and intervention areas, respectively, were enrolled in the study in November 2012, four months after the CPNP program had started. Enrolled children were followed over the next 12-months and measured for length and weight every three months. At each visit the mothers were asked about the child’s diet using 24-hour-based dietary recall questionnaire, and infant and young child feeding (IYCF) practices every month, whereas hand washing practices were elicited every six months. Length-for-age (LAZ), weight-for-length (WLZ), and weight-for-age (WAZ) z scores, and stunting (LAZ<-2), wasting (WLZ<-2), and underweight (WAZ<-2) status were calculated using the WHO reference population. Child feeding scores for current breastfeeding (range 0-7), meal frequency (range 0-15), dietary diversity (DDS, range 0-10), hand washing (range 0-6), and two composite feeding scores combining different variables were constructed to represent a 12-month feeding and hygiene practice pattern. For process evaluation, we reviewed attendance records for 3,299 participants and conducted group session observation checklists of 114 sessions and extracted data to construct process indicators such as dose delivered, dose received, and fidelity. Mothers were also asked every three months whether they were involved in the ENA and CMAM programs. A small survey of individuals who directly participated in the nutrition sessions (n = 197) was conducted to examine their own acceptance of the CPNP program and recall of messages provided in the nutrition sessions. Multilevel mixed-effects linear regression models using longitudinal data and Cox regression for recurrent event models for stunting, underweight, and wasting episodes were used. Results: We conducted an intention-to-treat analysis. A total of 1,475 children—750 and 725 children in the control and intervention area who were measured at least two out of five times, respectively—were included in the analysis of child growth. After controlling for clustering effects and enrollment characteristics, children in the intervention area had a significantly greater LAZ/month of 0.021 (95% CI: 0.008, 0.034), WAZ/month of 0.015 (95% CI: 0.003, 0.026), length/month of 0.059 cm (95% CI: 0.027, 0.092), and weight/month of 0.031kg (95% CI: 0.019, 0.042) compared to children in the control area. There was no difference in the risk of recurrent incidence of stunting [Hazard Ratio (HR) 1.01, 95% CI: 0.811], underweight (HR 0.826, 95% CI: 0.643, 1.061), or wasting (HR 1.048, 95% CI: 0.812, 1.354) between intervention and control areas. Prevalence of stunting was lower in the intervention area by 7.5% at the 9 month follow-up (p = 0.037) and by 7.3% at 12 month of follow-up (not significant). Of 1,199 mothers, those in the intervention area (n = 570) showed higher scores on meal frequency (diff: 1.02, 95% CI: 0.33, 1.70), and both composite feeding scores (type 1, diff: 1.30, 95% CI: 0.41, 2.19) (type 2, diff: 1.39, 95% CI: 0.48, 2.23) than mothers in the control area (n = 629), but there were no differences in scores of current breastfeeding (diff: -0.08, 95% CI: -0.22, 0.06), dietary diversity (diff: 0.38, 95% CI: -0.21, 0.98), or hand washing (diff: 0.41, 95% CI: -0.21, 1.02) between mothers in the two areas. The CPNP sessions were delivered at 81% of the expectation level; fidelity to intervention protocols varied between 68% and 92%. Out of 197 CPNP participants, 95% had ever contributed materials to a session and 90% showed positive perceptions of the session practices. For each of the session activities (e.g., food preparation), ~70% rated their participation level as active. The participants recalled 5.0 ± 2.0 (SD) out of the 8 messages (e.g., preparing nutritious complementary foods). We also observed improvements in feeding and hygiene, i.e., minimum dietary diversity (34.0% vs. 19.9%, p < 0.05) and mother’s cleanliness (40.0% vs. 24.0%, p < 0.05) at an early stage of the CPNP, and a higher involvement of CPNP participants in the existing program (e.g., Essential Nutrition Action) over the project period compared to non-participants within the same intervention area (32.3% vs. 19.8%, p < 0.001). Conclusions: A community-based participatory nutrition program approach adapting the positive deviance model was found to be effective in improving child feeding practices among caregivers and child growth in the rural Ethiopian context. Using a PIP analysis, our results suggest that the CPNP program was adequately implemented and well-accepted among participants. Program impact was seen in the early stages of the program implementation, although a true baseline was not available, giving the additional benefit of positively influencing the ongoing ENA program, and eventually resulting in improved child nutritional status.
Description
Keywords
Child undernutrition, Community-based participatory nutrition promotion, cluster randomized trial
Citation