Study Participants
The mean ± SD age of the mother was 25.5 ± 4.7 years: the majority (71%) were 21-30 years. About half the women were in primary school or had completed primary school. The mean ± SD age of young children was 9 ± 2.6 months, with about half of them between 6-8 months. A summary of the study participants' socio-demographic characteristics is presented in Table 1.
Table 1. Socio-Demographic Characteristics of Study Participants at Baseline, Comparison of the Intervention Group (IG) and Control Group (CG), Southern Ethiopia, 2016 (n=771)
Socio-demographic characteristics
|
IG n (%)
|
CG n (%)
|
P
|
Age of the mother (years)
< 24 25-34
>35
|
146 (37.8)
227 (58.8)
13 (3.4)
|
175 (45.5)
207 (53.8)
3 (0.8)
|
0.59
|
Marital status
Married
Divorced
Widowed
|
365 (94.6)
14 (3.6)
7 (1.8)
|
379 (98.4)
4 (1.0)
2 (0.5)
|
0.004
|
Educational status
Illiterate
Read and write
|
150 (38.9)
236 (61.2)
|
147 (38.2)
238 (61.8)
|
0.58
|
In charge of food purchase
Yes No
|
150 (38.9)
236 (61.1)
|
158 (59.0)
158 (41.0)
|
0.54
|
Source of income generating activities by woman
Yes (petty trade, day labor) No
|
62 (16.1)
324 (83.9)
|
49 (12.7)
336 (87.3)
|
0.19
|
Household size
Low (1-4 family members) Medium (5-10 family members)
Large (>10 family members)
|
97 (25.1)
222 (57.5)
67 (17.4)
|
124 (32.2)
190 (49.4)
71 (18.4)
|
0.45
|
Age of the children (months)
6-8 9-11
12-15
|
172 (44.6)
123 (31.9)
91 (23.6)
|
197 (51.2)
110 (28.6)
78 (20.3)
|
0.19
|
Sex of children
Male Female
|
211 (54.7)
175 (45.3)
|
201 (52.2)
184 (47.8)
|
0.50
|
Wealth Index
Poor
Better
|
255 (66.1)
131 (33.9)
|
229 (59.5)
156 (40.5)
|
0.55
|
Household hunger
No household hunger
Mild household hunger
Moderate household hunger
Severe household hunger
|
77 (19.9)
46 (11.9)
125 (32.4)
138 (35.8)
|
140 (36.4)
27 (7.0)
84 (21.8)
134 (34.8)
|
0.10
|
P is significant at <0.05.
Knowledge, Attitude, and Practices of Mothers
Both the intervention and control group mothers had low scores on KAP at baseline. At midline, KAP improved in the intervention group (Figure 2). After the 9-month nutrition education, mothers' mean knowledge (p=0.001) and practices (p=0.001) significantly improved, but the attitude score remained the same as the midline score. There was a significant main effect of nutrition education on knowledge (F= 488.498; df = 2; p=0.001), attitude (F= 375.221; df=2; p=0.001), and practices (F= 201.431; df=2; p=0.001) within groups. A similar significant effect was seen between groups on knowledge (F=3071.99; df=1; p=0.001), attitude (F= 1297.50; df= 1; p=0.001) and, practices (F=158.98; df=1; p=0.001).
Household Processing of Pulses and Preparation of Complementary Foods by the Mothers
At baseline, none of the mothers from either the intervention or control group indicated having ever soaked and germinated pulse products. At midline, 132 (43%) from the intervention and 29 (9.2%) from the control group reported that they had soaked and germinated pulses. Only 13 (4.2%) from the intervention group could recall from memory all the steps in household processing techniques (sorting, soaking, draining and soaking for 48 h, sun drying, roasting in a warm pan, and milling), and no members of the control group could describe all of these steps. In addition, only five women (1.6%) from the intervention and three (1%) from the control group were able to recall the right proportion of cereal-pulse mix, which is 3/4th cereal and 1/4th pulse. At endline, 214 (69.7%) from the intervention and 37 (11.8%) from the control group reported that they soak and germinate pulse crops; only 23 (7.5%) from the intervention group could recall from memory all six steps of the household processing techniques, while none of the control group could. One hundred and forty-six (47.6%) from the intervention and 15 (4.8%) from the control group had learned the right proportion of cereal-pulse mix for preparing complementary food (Table 2).
Table 2. Percentage of Selected Practice Variables of Intervention Group (IG) and Control Group (CG), Southern Ethiopia, 2016 (IG n=307; CG n=314)
Variables
|
Midline
|
X2
|
Endline
|
X2
|
N
|
%
|
|
N
|
%
|
|
Prepare pulse incorporating CF
IG
CG
|
289
211
|
94
67.2
|
0.001
|
302
213
|
98.4
67.7
|
0.001
|
Soak and germinate pulse crops
IG
CG
|
132
29
|
43.0
9.2
|
0.001
|
414
37
|
69.7
11.8
|
0.001
|
Show 3 or more steps of household processing techniques
IG
CG
|
88
1
|
28.6
0.3
|
0.001
|
117
3
|
38
0.9
|
0.001
|
Mix cereal and pulse crops in the right proportions
IG
CG
|
5
3
|
1.6
1.0
|
0.001
|
146
15
|
47.6
4.8
|
0.001
|
Midline: 4.5 months; Endline: 9.0 months
Pulse Consumption, Dietary Diversity, Meal Frequency, and Nutrient Intake
Haricot bean was the most commonly used pulse crop in the study area. Mothers who reported using pulses once or more than once per day in complementary food increased from 11.1% at baseline to 80.1% at endline in the intervention group and from 15.3% at baseline to 58.6% at endline in the control group. At baseline, there was no significant difference between the intervention and control groups in children's consumption of pulses (p=0.47). At midline and endline, there was a significant difference in consumption of pulses between the control and intervention groups (p=0.001). Figure 3 shows the frequency of pulse consumption by young children in both the intervention and control groups.
Mean dietary diversity at baseline was similar for the intervention children (2.1 ± 1.0) and controls (2.2 ±0.8); at midline, values were 2.5±0.8 and 2.2 ±0.9, respectively and, at endline, 3.7 ± 1.4 and 3.2 ±1.4, respectively. At baseline, only 23 (7.5%) from the intervention and 12 (3.8%) from the control group had consumed food from four or more food groups in the last 24 hours prior to data collection. At midline, 27 (8.8%) from the intervention and 26 (8.3%) from the control group had consumed food from four or more food groups in the same period. At endline, 158 (51.5%) from the intervention and 136 (43.3%) from the control had consumed food from four or more food groups. There was a statistically significant difference between baseline and midline diet diversity (p=0.001) and between midline and endline (p=0.001) between and within groups, which indicated change overtime in both groups. Table 3 summarizes the food groups consumed by young children in the 24 hours prior to data collection at baseline, midline, and endline.
Table 3 Food Groups of Young Children in the Last 24 hours Prior to Data Collection Date at Baseline, Midline (4.5 months) and Endline (9 months), Southern Ethiopia, 2016 (Intervention Group, IG n=307; Control group, CG n=314)
Food group
|
Baseline
N (%)
|
Midline
N (%)
|
Endline
N (%)
|
|
IG
|
CG
|
IG
|
CG
|
IG
|
CG
|
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
Cereal & root crops
|
236
(76.9)
|
273
(86.9)
|
281
(91.5)
|
281
(89.5)
|
304
(99)
|
312
(99.4)
|
Pulse & nuts
|
96
(31.3)
|
121
(38.5)
|
189
(61.1)
|
135
(43)
|
264
(86)
|
81
(25.8)
|
Milk & milk group
|
27
(8.8)
|
12
(3.8)
|
31
(10.1)
|
10
(3.2)
|
134
(43.6)
|
128
(40.8)
|
Meat & organ meat
|
0
|
0
|
0
|
0
|
4
(1.3)
|
6
(1.9)
|
Egg
|
27
(8.8)
|
12
(3.8)
|
42
(13.7)
|
64
(20.4)
|
147
(47.9)
|
168
(53.5)
|
Vitamin A rich fruits and vegetables
|
0
|
0
|
105
(34.2)
|
74
(23.6)
|
192
(62.5)
|
198
(63.1)
|
Other fruits and vegetables
|
265
(86.3)
|
283
(90.1)
|
116
(37.8)
|
133
(42.4)
|
106
(34.5)
|
99
(31.5)
|
Mean Diet Diversity Score
|
2.1 ±1.0
|
2.2 ±0.8
|
2.5 ±0.8
|
2.2 ±0.9
|
3.7 ±1.4
|
3.2 ±1.4
|
IG: Intervention Group; CG: Control Group
Nutritional Status of Children
Anthropometric measurements were measured in the study children at baseline, midline, and endline. At baseline, there was no significant difference between the intervention and control groups in all anthropometry measurements and anthropometry indices. At endline, the mean age of male children was 18.8 ± 2.9 mo and female children was18.7 ± 2.9 mo. After intervention there was a significant difference (p < 0.05) in all anthropometric indices in the intervention group compared to the control group, as well as differences over time (Table 4). At baseline, low height and weight measurements were reflected in the high prevalence of stunting, wasting, and underweight. At midline and endline measurements, there were increases in the prevalence of stunting in both intervention and control groups; however, wasting and underweight improved only in the intervention group.
Table 4 Anthropometric Status of Children at Baseline, Midline (4.5 months) and Endline (9 months), IG n=307; Control group, CG n=314)at Sidama Zone, Southern Ethiopia, 2016
Anthropometry measurements
|
Mean (SD) (N=771)
|
|
Mean (SD) (N=621)
|
|
Baseline
|
Midline
|
Endline
|
Weight (kg)
Intervention
Control
Length (cm)
Intervention
Control
MUAC (cm)
Intervention
Control
|
8.3 (±1.3)a
7.3 (±1.2)a
69.3 (± 4.7)a
68.9 (±1.1)
13. 6 (±1.1)a
13.6 (±1.0)a
|
|
9.3b,x
8.7b,y
74.3b,x
73.6b,x
13.8b,x
13.4b,y
|
10.2c,x
9.4c,y
78.1c,x
77.1c,y
14.4c,x
13.9c,y
|
Anthropometric indices
|
Wasting (weight-for-length z score)
Intervention
Control
|
-0.44 (±1.2)a
-0.73 (±1.1)a
|
|
0.06b,x
-0.5b,y
|
0.28c,x
-0.43c,y
|
Stunting (length-for-age z score)
Intervention
Control
|
-1.05 (±1.4)a
-1.10 (±1.2)a
|
|
-1.2b,x
1.3b.x
|
-1.5c,x
-1.7c,y
|
Underweight (weight-for-age z score)
Intervention
Control
|
0.23 (±1.2)a
-0.13 (±1.1)a
|
|
-0.5b,x
-1.3b,y
|
-0.5c,x
-1.2c,y
|
MUAC z score
Intervention
Control
|
-0.68 (±1.0)a
-0.61 (±1.0)a
|
|
-0.7b,x
-1.0b,y
|
-0.4c,x
-0.7c,y
|
|
|
|
|
|
|
Note: SD= Standard Deviation; Values in rows are significantly different (p<0.05) from the baseline if they have the letter "b" or "c"; values with the letter "c" are significantly different from values having "b." Values for intervention and control in the midline and endline columns are significantly different if they have the letter "x" or "y"; values with the letter "x" are significantly different from values of "y" (p<0.05).
MUAC: mid-upper arm circumference.
Differences between the intervention and control groups were largely seen in all three measures: stunting, wasting, and underweight during the course of the study. At endline, the prevalence of stunting increased more in the control group than in the intervention group, and the difference was significant (p=0.02). At baseline, 14.4% in the intervention group and 25.8% in the control group were wasted, and prevalences did not change much at midline (15.3% and 28.1%, respectively); at endline, the prevalence of wasting had significantly increased (p=0.001) in the control group (29.7%) over time, while the intervention group (11.8%) had declined from baseline, showing a significant positive effect of the intervention. At baseline, 32.9% of children in the intervention group and 34% of children in the control group were underweight; at midline only the control group showed a worsening (47.4% of children). Both groups had a decreased prevalence of underweight (11.7% and 29.3%, respectively) at endline, but more children were underweight in the control group (p=0.001). These changes in the children in the two groups were reflected in MUAC measurements. At baseline, 35 % of children from the intervention group and 37% of children from the control group had lower MUACZ-scores. Although the number of children who had lower MUACZ scores increased in both the intervention (40%) and control group (54%), at midline, more children in the control group had lower MUACZ-scores compared to the intervention group. At endline, the number of children with low MUACZ scores had decreased in both the intervention (28%) and control (38.6%) groups compared to midline. However, at endline the number of children who had low MUACZ-scores was greater in the control group than in the intervention group (p=0.001) (Table 4).
Women's Empowerment in Household Decision Making
All participating mothers (n=386) in the intervention group reported that they had planted the haricot bean seed provided to them after the midline data collection. The seeds were later harvested following the endline data collection. Several visits to the field were carried out by the agriculture agents and the project team to a selected group of women; however, we did not collect information on how much they harvested, how much they kept for consumption, or how much they sold during endline data collection.