Table 1 provides insights into the associations between demographic and socioeconomic factors and the prevalence of stunting among children at baseline. It is evident that there is a notable difference in stunting between male and female children. Across the categories of moderately and severely stunted, male children tended to have higher proportions than did their female counterparts. For instance, in the severely stunted category, 62.66% of male children were affected, while 37.34% of female children were affected. In terms of place of residence, rural areas had greater proportions of stunted children than did urban areas. Regional differences are also notable, with the Tigray, Amhara, and Oromiya regions exhibiting relatively greater stunting than southern nationalities and peoples (SNNPs) and Addis Ababa.
Table 1
Background characteristics of children, mothers, and households across stunting levels in 2002 in Ethiopia (n = 1,536).
Variables | Categories | Stunting levels |
Not stunted | Moderately stunted | Severely stunted |
Sex of child’s | Male | 443(49.17%) | 176(55.17%) | 198(62.66%) |
Female | 458(50.83%) | 143(44.82%) | 118(37.34%) |
Place of residence of child’s | Urban | 337(37.40%) | 94(29.47%) | 72(22.78%) |
Rural | 564(62.60%) | 225(70.53%) | 244(77.22%) |
Religion of child’s | Muslim | 177(19.64%) | 46(14.42%) | 49(15.51%) |
Protestant | 96(10.65%) | 33(10.34%) | 45(14.24%) |
Orthodox | 612(67.93%) | 232(72.73%) | 218(68.99%) |
Other | 16(1.78%) | 8(2.51%) | 4(1.26%) |
Region | Tigray | 198(21.98%) | 72(22.57%) | 53(16.77%) |
Amhara | 139(15.43%) | 79(24.77%) | 84(26.58%) |
Oromiya | 171(18.97%) | 71(22.26%) | 74(23.42%) |
SNNP | 243(26.97%) | 65(20.37%) | 79(25.00%) |
Addis Ababa | 150(16.65%) | 32(10.03%) | 26(8.23%) |
Mother’s education level | None | 506(56.16% | 213(66.77%) | 227(71.84%) |
Primary school | 303(33.63% | 83(26.02%) | 77(24.37%) |
Secondary school and above | 86(9.55%) | 19(5.96%) | 8(2.53%) |
other | 6(0.66%) | 4(1.25%) | 4(1.26%) |
Father’s education level | None | 397(44.06%) | 158(49.53%) | 180(56.96%) |
Primary education | 285(31.64%) | 88(27.59%) | 76(24.05%) |
Secondary education and above | 128(14.21%) | 21(6.58%) | 23(7.28%) |
other | 91(10.09%) | 52(16.30%) | 37(11.71%) |
Access to safe drinking water | Yes | 495(54.94% | 168(52.66%) | 136(43.04%) |
No | 406(45.06%) | 151(47.34%) | 180(56.96%) |
Access to sanitation | Yes | 389(43.17%) | 96(30.09%) | 66(20.89%) |
No | 512(56.83%) | 223(69.91%) | 250(79.11%) |
Access to adequate fuels cooking | Yes | 80(8.88%) | 22(6.89%) | 15(4.74%) |
No | 821(91.12%) | 297(93.10% | 301(95.25%) |
With respect to religion, there are variations in stunting levels across different religious groups, with Orthodox followers showing the highest percentage of stunted children. The education levels of both mothers and fathers seemed to correlate inversely with stunting levels, with higher education levels associated with lower percentages of stunted children. Additionally, access to basic necessities such as safe drinking water, sanitation, and adequate cooking fuels appears to have an impact on stunting, with no access corresponding to higher stunting levels.
Figure 1 presents significant insights into the trends in stunting levels, namely, nonstunted, moderately stunted, and severely stunted, from 2002 to 2016. The proportion of nonstunted children, represented by the blue line, peaks in 2009 and then shows a slight decline, consistently comprising the largest group. This suggests initial improvements in factors contributing to child growth, such as nutrition and healthcare, although the decline after 2009 indicates potential challenges in sustaining these improvements.
The proportion of moderately stunted children, depicted by the green line, exhibited minor fluctuations without a clear trend. This finding suggests that while some progress has been made in addressing moderate stunting, it has not been consistent over the years.
Severely stunted children, indicated by the red line, remain consistently low and stable, suggesting that severe stunting, though less prevalent, continues to affect a small but persistent portion of the population. This persistence indicates ongoing issues that prevent the complete eradication of severe stunting, possibly due to entrenched socioeconomic factors, persistent poverty, or inadequate access to critical resources.
Overall, the data suggest that while there has been progress in reducing stunting, particularly in the nonstunted category, efforts need to be intensified and sustained to effectively address moderate and severe stunting. The persistence of stunting at moderate and severe levels underscores the need for comprehensive and long-term interventions targeting the underlying causes.
Table 2
GEE ordinal logistic regression model parameter estimates, standard errors and p values.
Predictors | Categories | Estimate | san.se | san.z | P - value | |
Intercept 1 | | 0.2315 | 0.2388 | 0.9692 | 0.3324 | |
Intercept 2 | | 1.6538 | 0.2394 | 6.9076 | 0.0000 | *** |
Time | 2002 (ref) | | | | | |
2006 | 1.9480 | 0.4930 | 3.9517 | 0.0001 | *** |
2009 | 3.3528 | 0.8351 | 4.0148 | 0.0001 | *** |
2013 | 4.3585 | 1.2942 | 3.3676 | 0.0008 | *** |
2016 | 5.3631 | 1.6402 | 3.2697 | 0.0011 | ** |
Place of residence | Urban (ref) | | | | | |
Rural | -0.0770 | 0.1104 | -0.6971 | 0.0486 | * |
Sex of children | Male (ref) | | | | | |
| Female | 0.4988 | 0.0815 | 6.1216 | 0.0000 | *** |
Religion of children | Muslim (ref) | | | | | |
| Protestant | -0.3021 | 0.1613 | -1.8730 | 0.0611 | . |
| Orthodox | 0.0327 | 0.1367 | 0.2389 | 0.8112 | |
| Other | -0.1224 | 0.3000 | -0.4081 | 0.6832 | |
Age in months | | -0.0279 | 0.0096 | -2.8968 | 0.0038 | ** |
Household size | | -0.0085 | 0.0140 | -0.6090 | 0.5425 | |
Wealth index | | 1.0308 | 0.2903 | 3.5505 | 0.0004 | *** |
Access to safe drinking water | No (ref) | | | | | |
Yes | -0.1389 | 0.0575 | -2.4143 | 0.0158 | * |
Access to sanitation | No (ref) | | | | | |
Yes | 0.0381 | 0.0561 | 0.6796 | 0.4967 | |
Access to adequate fuels cooking | No (ref) | | | | | |
Yes | -0.4299 | 0.1122 | -3.8302 | 0.0001 | |
Mothers education level | None (ref) | | | | | |
Primary school | 0.2188 | 0.0897 | 2.4402 | 0.0147 | *** |
Secondary school and above | 0.4420 | 0.1725 | 2.5629 | 0.0104 | * |
Other | -0.0873 | 0.1134 | -0.7700 | 0.4413 | |
Father’s education level | None (ref) | | | | | |
Primary school | 0.0603 | 0.0831 | 0.7254 | 0.4682 | |
Secondary school and above | 0.3359 | 0.1566 | 2.1455 | 0.0319 | * |
Other | -0.1206 | 0.0838 | -1.4390 | 0.1502 | * |
Region | Tigray(ref) | | | | | |
Amhara | -0.2980 | 0.1261 | -2.3638 | 0.0181 | ** |
Oromiya | -0.2322 | 0.1293 | -1.7963 | 0.0725 | * |
SNNP | 0.0363 | 0.1618 | 0.2246 | 0.8223 | . |
Addis Ababa | 0.3878 | 0.1903 | 2.0377 | 0.0416 | * |
Signif. Codes: 0 ‘***’ 0.001 ‘**’ 0.01’*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
As shown in Table 2, the GEE ordinal logistic regression model revealed several significant predictors of child stunting, categorized as not stunted, moderately stunted, and severely stunted. The time trends indicate that children measured in 2006, 2009, 2013, and 2016 had significantly greater odds of being less stunted than those measured in 2002. Specifically, the odds ratios (ORs) for these years were 7.01, 28.56, 77.80, and 212.48, respectively, with all p values < 0.001 except for 2016, for which the p value was < 0.01.
Place of residence was also a significant predictor, with children in rural areas having slightly lower odds of being less stunted than their urban counterparts (OR = 0.93, p < 0.05). Additionally, the sex of the children was a strong predictor, with female children having significantly greater odds of being less stunted than male children (OR = 1.65, p < 0.001).
Religious affiliation had some impact, with Protestant children having a trend toward lower odds of being less stunted than Muslim children, although this finding was marginally significant (OR = 0.74, p < 0.1). No significant differences were observed for Orthodox or other religions.
Age in months negatively influenced the odds of being less stunted (OR = 0.97, p < 0.01), indicating that older children were more likely to be stunted. Household size did not have a significant effect on stunting outcomes.
Socioeconomic factors played a crucial role, with a higher wealth index significantly increasing the odds of children being less stunted (OR = 2.80, p < 0.001). Access to safe drinking water slightly decreased the odds of being less stunted (OR = 0.87, p < 0.05), while access to adequate fuels for cooking significantly decreased the odds (OR = 0.65, p < 0.001).
Parental education levels were also significant. Mothers with primary education had increased odds of their children being less stunted (OR = 1.24, p < 0.05), and mothers with secondary education or higher had an even greater effect (OR = 1.56, p < 0.05). For fathers, only secondary education and above had a significant positive impact (OR = 1.40, p < 0.05).
The regional differences were notable, with children from Amhara having lower odds of being stunted than those from Tigray (OR = 0.74, p < 0.05). Oromiya showed a marginally significant trend towards lower odds (OR = 0.79, p < 0.1), while there was no significant difference for SNNP. Conversely, children from Addis Ababa had greater odds of being less stunted (OR = 1.47, p < 0.05).
Overall, the findings highlight the importance of temporal trends, place of residence, gender, age, socioeconomic status, access to resources, parental education, and regional differences in influencing child stunting outcomes.