Effect of Behavioral Change Communication and Livestock Feed Intervention on Dietary Practices in a Kenyan Pastoral Community: A Randomized Controlled Trial

Abstract

Low dietary diversity is a key driver of undernutrition and remains a significant public health challenge in low- and middle-income countries. This study evaluated the effect of nutritional counselling and the provision of livestock feed, aimed at sustaining milk production during dry periods, on the dietary diversity of women and children in a pastoralist setting. Methods: A cluster randomized controlled trial was conducted among households in Laisamis subcounty, north-eastern Kenya, which were assigned to one of three arms: (1) an intervention arm providing livestock feed during critically dry periods, (2) an intervention arm providing livestock feed plus enhanced nutritional counselling (provided once a week, covering topics including hygiene and sanitation, breastfeeding, maternal nutrition, immunization and complementary feeding) or (3) a control arm. The dietary diversity of mothers and children was assessed every six weeks over two years. Panel difference-in-difference regression models were used to estimate intervention effects on dietary outcomes including child minimum dietary diversity (MDD), minimum acceptable diet (MAD), women's dietary diversity (MDD-W) and food security. Results: A total of 1734 households participated (639 in arm 1, 585 in arm 2, and 510 in the control arm). The provision of livestock feed alone had significant gains in child MAD (OR 1.20; 95% CI 1.08-1.34), child MDD (OR 1.15; 1.11-1.20), and MDD-W (OR 1.10; 1.01-1.19) whereas combined livestock feed with counselling, reduced child food poverty (OR 0.89; 95% CI 0.81-0.99), increased child MAD (OR 1.39; 1.22-1.52), and improved MDD-W (OR 1.21; 1.16-1.28) relative to control. Neither intervention increased child minimum meal frequency relative to control. Purchasing livestock was associated with higher odds of meeting dietary-diversity indicators but a lower meal frequency (OR 0.80; 0.80-0.90); in contrast, cash-transfer receipt was linked to reduced odds of achieving child MDD (OR 0.90; 0.87-0.94), child MAD (OR 0.95; 0.85-0.97), and women's MDD (OR 0.73; 0.54-0.89). Conclusions: Livestock feed provision sustains milk consumption and improves dietary diversity in pastoralist populations. When combined with nutritional counselling, these interventions strengthen the link between animal and human health, with important implications for food security.

Keywords: arid and semi-arid lands; dietary diversity; livestock for human health.

Figure A1

Likelihood of monthly change in the indicators of food groups over time for children aged 6–23 months in the three study arms (animal feed intervention, animal feed and enhanced counselling, and control arm). The grey shaded areas represent dry seasons.

Figure A2

Likelihood of monthly change in the indicators of food groups over time for children aged 24–59 months in the three study arms (animal feed intervention, animal feed and enhanced counselling, and control arm). The grey shaded areas represent dry seasons.

Figure A3

Likelihood of monthly change in the indicators of food groups over time for mothers of reproductive age in the three study arms (animal feed intervention, animal feed and enhanced counselling, and control arm). The grey shaded areas represent dry seasons.

Figure A4

Likelihood of overall change in the indicators of dietary diversity scores. The vertical red dashed line marks the null value (OR = 1); estimates whose 95% CIs do not cross this line are statistically significant.

Figure 1

Map of Laisamis subcounty showing the selected sublocations and the villages in the different study arms. The inset map on the left shows the location of Marsabit County.

Figure 2

Overview of variables collected during each collection period.

Figure 3

Participant flow diagram for the cluster randomized controlled trial.

Figure 4

Food groups consumed by children <5 years over the study period. The dots represent the actual monthly proportion for each food group, whereas the line represents the trend line.

Figure 5

Proportion of women in the different intervention arms consuming each of the different food groups over the study period. The dry seasons occurred from January to March and July to October. Season 1 corresponds to September to December 2019, and Season 9 corresponds to November to December 2021.

Figure 6

Likelihood of overall change in the indicators of dietary diversity score (a) and seasonal change over time (b). The dry seasons occurred from January to March and July to October. Season 1 corresponds to September to December 2019, and Season 9 corresponds to November to December 2021.