Effectiveness of Microfinance, Nutrition Education, and Couples Strengthening on Maternal and Child Diets in Rural Niger: Evidence From a Program Evaluation Project
Restricted (Penn State Only)
- Author:
- Shakya, Evaniya
- Graduate Program:
- Nutritional Sciences
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- June 14, 2024
- Committee Members:
- Kristina Petersen, Professor in Charge/Director of Graduate Studies
Stephen Kodish, Major Field Member
Muzi Na, Chair & Dissertation Advisor
Alison Gernand, Major Field Member
Stephanie Lanza, Outside Unit & Field Member - Keywords:
- Niger
Dietary diversity
Program evaluation - Abstract:
- The majority of Nigerien women and children have inadequate diets, which is an immediate driver of undernutrition. Only 6.3% of children aged 6–23 months and 37% of women consumed a minimally diverse diet (≥ five food groups) in Niger. Some key barriers to improving maternal and child diets in rural Niger include frequent environmental and economic shocks, low education levels, and inequitable gender and social norms. Presently, we lack evidence on the effectiveness of multicomponent programs targeting these complex challenges. This dissertation aims to fill this significant gap in the literature by evaluating the impact of interventions encompassing microfinance, nutrition education, and couples strengthening strategies on dietary diversity among mothers and children aged 6–59 months in rural Niger. The overall study aimed to evaluate three village-level interventions implemented by Catholic Relief Services (CRS)-led Girma project (2018-2025) to improve maternal and child diets in Zinder, Niger. These interventions were a savings-led microfinance approach called Savings and Internal Lending Communities (SILC), nutrition education for mothers using a Care Group model (CARE), and a nutrition-integrated couples strengthening approach called Maison Familiale Harmonieuse (MFH). Between February and March 2022, we enrolled 2,331 households with a husband, a wife aged 15–49 years, and a child aged 6–59 months in a prospective quasi-experimental study, and a baseline survey was conducted. We compared three study arms: Intervention Group 1 (I1, SILC and CARE), Intervention Group 2 (I2, SILC, CARE, and MFH), and control group (no intervention). The follow-up endline survey occurred after 12 months, between February–March 2023. At both timepoints, 24-hour dietary recalls were conducted. Child dietary diversity scores (CDDS) for children aged 6–59 months were measured based on World Health Organization (WHO) guidelines. Women's dietary diversity scores (WDDS) were measured based on Food and Agriculture Organization (FAO) guidelines. The objective of the first study was to examine the effectiveness of I1, I2, and control, in improving maternal dietary diversity. Mothers aged 15–49 years of children aged 6–59 months (ncontrol= 409, nI1 = 655, nI2 = 910) with valid WDDS at both timepoints were included in the study. To examine the difference in WDDS from baseline to endline (WDDSdiff) between the three study arms, we used multivariate regression models with cluster robust standard errors (SE). The baseline characteristics differing between all study arms were included as covariates in the regression models. Additionally, using modifier-by-study arm interactions, this study explored whether maternal characteristics (marriage structure, age, and education) modified the intervention effects. We found that WDDS increased significantly in I1 (WDDSdiff: 0.43, 95% CI: 0.22, 0.64) and I2 (WDDSdiff: 0.52, 95% CI: 0.31,0.73), but not in control (WDDSdiff: 0.03, 95% CI: -0.22, 0.29). The increase in WDDS was significantly higher in I1 (0.40, 95% CI: 0.15, 0.65) and I2 (0.48, 95% CI: 0.27, 0.70) compared to control. However, no significant difference in WDDSdiff was observed between I1 and I2. Maternal education was the only significant effect modifier of the intervention effects (p-interaction = 0.032). The objective of the second study was to assess the effects of I1, I2, and control in improving CDDS among children aged 6–59 months. The analytic sample of the study included children 6–59 months (ncontrol= 387, nI1 = 620, nI2 = 880) with valid CDDS at both timepoints. First, the difference in CDDS between endline and baseline (CDDSdiff) was calculated. Then, pairwise differences in CDDSdiff between control, I1, and I2 were examined using adjusted multivariate regression with cluster robust SEs. Modifier-by-study arm interactions examined maternal and child characteristics (marriage structure, maternal age, maternal education, child age, and child sex) as potential effect modifiers. Finally, we examined percentage point change in consumption of food groups by study arm and child age strata. We observed an increase in CDDS from baseline to endline in control (CDDSdiff: 0.49, 95% CI: 0.22, 0.75), I1 (CDDSdiff: 0.55, 95% CI: 0.37, 0.74), and I2 (CDDSdiff: 0.75, 95% CI: 0.53, 0.96). However, no significant intervention effect was seen between any of the study arms. Only baseline child age modified these effects (p-interaction = 0.022). Among children aged 12–23 months, CDDS improved more with both interventions compared to the control. Among 24–59-month-olds, CDDS improved more in I2 than in I1 or control. Overall, the consumption of animal-source foods was very low. Observing the percentage point changes, the consumption of flesh foods and eggs increased very little while the consumption of dairy decreased. The objective of the third study was to investigate psychosocial pathways linking SILC, CARE, and MFH interventions to improved WDDS. We combined I1 and I2 into one intervention group in this study as no difference in intervention impact was seen between the two groups. This analysis included couples with WDDS and psychosocial scores at both timepoints (ncontrol=329, nintervention= 1,401). We used structural equation modeling to examine 1) the mediation of the intervention effect on WDDS by wives’ and husbands’ intentions towards maternal nutrition and microfinance and 2) the distal psychosocial pathways (knowledge, attitude, self-efficacy, and subjective norms) linking interventions to intentions. All models were adjusted for baseline characteristics differing between study arms and village-level cluster robust SEs. Among wives, intentions towards maternal nutrition [β (SE): 0.28 (0.06), p <0.001] and microfinance [β (SE): 0.18 (0.06), p: 0.004] increased in the intervention group compared to the control group, but not among husbands. All indirect effects through wives’ distal psychosocial pathways leading to their intentions toward maternal nutrition were non-significant. Unexpectedly, wives’ microfinance-related self-efficacy had a negative indirect effect on their intentions toward budgeting for nutritious foods. We observed that the intervention effects on husbands’ psychosocial scores were either negative or not significant. Our findings provide evidence that combined SILC and CARE, with or without MFH, improved maternal but not child diets. Findings from this research also indicate that future strategies should target the nutritional needs of children in different age groups. Overall, the improvement in maternal dietary diversity was not mediated by the intentions of wives or husbands toward improving maternal diets or budgeting for nutritious foods. Further studies should target synergistic improvements in couples’ psychosocial constructs related to maternal and child nutrition. In conclusion, continuous efforts are needed to address the suboptimal diets of mothers and children in Niger.