Determinants of Improved Household Sanitation Use in Ethiopia: A Multilevel Logistic Regression Analysis

ABSTRACT

Introduction

Sanitation is a fundamental human right and a cornerstone of public health. In Ethiopia, access to improved sanitation facilities remains limited, especially in rural areas. Poor sanitation, coupled with inadequate water supply and hygiene, significantly contributes to illness and mortality worldwide, disproportionately affecting developing countries. This study aims to identify individual and community‐level factors associated with the use of improved sanitation facilities among Ethiopian households by applying a multilevel logistic regression model to data from the 2019 Ethiopian Mini Demographic and Health Survey (EMDHS).

Methods

This study used data from the 2019 Ethiopian Mini Demographic and Health Survey (EMDHS. A total of 6141 households were included in the analysis. A cross‐sectional design was used to estimate the use of improved sanitation services at national, regional, urban, and rural levels. To account for the hierarchical nature of the data households nested within community's multilevel logistic regression was employed. This approach allowed for the partitioning of variance into individual and community level components, thereby improving model accuracy and capturing contextual influences on sanitation use.

Results

The analysis revealed significant regional disparities in the use of improved sanitation services among Ethiopian households. Married households had higher odds of using improved sanitation (AOR = 1.81, 95% CI: 1.67–4.70), while unmarried (AOR = 0.35, 95% CI: 0.21–0.56) and divorced households (AOR = 0.24, 95% CI: 0.23–0.50) had lower odds compared to widowed households. Rural households were less likely to use improved sanitation than urban ones (AOR = 0.37, 95% CI: 0.140–0.99). Illiterate (AOR = 0.13, 95% CI: 0.08–0.20) and primary‐educated (AOR = 0.25, 95% CI: 0.16–0.39) household heads were less likely to access improved sanitation than those with higher education. Poor (AOR = 0.86, 95% CI: 0.44–0.98) and middle‐income (AOR = 0.84, 95% CI: 0.42–0.93) households had lower odds compared to rich households. Homeownership (AOR = 2.28, 95% CI: 1.13–2.46) and private latrine use (AOR = 1.43, 95% CI: 1.36–2.65) were significantly associated with higher sanitation use.

Conclusion

This study reveals substantial regional and sociodemographic disparities in the use of improved sanitation facilities in Ethiopia. Households headed by married, educated, and wealthier individuals living in urban areas with private latrines and home ownership are more likely to have improved sanitation. Targeted interventions focusing on rural, poorer, less educated, and female‐headed households are essential to enhance access to improved sanitation. Addressing these disparities is critical for Ethiopia to achieve its national sanitation goals and contribute toward the Sustainable Development Goals related to health and well‐being.

Abbreviations

AIC

akaike information criteria

BIC

Bayesian information criteria

FMoH

Federal Ministry of Health

HH

household

ICC

intra‐class correlation

OR

odds ratio

SNNPR

South Nationality of People Region

1. Introduction

1.1. Background of the Study

Sanitation, the safe disposal of human waste and promotion of hygiene is fundamental to public health and disease prevention [1]. Globally, inadequate sanitation contributes to significant morbidity and mortality, particularly from diarrheal diseases, which disproportionately affect children under five in low‐income countries like Ethiopia [2]. According to the World Health Organization, poor sanitation, coupled with unsafe water and hygiene practices, is a leading cause of diarrheal illness responsible for approximately 1.8 million deaths annually, with 90% of these occurring in developing regions [3, 4, 5, 6].

In Ethiopia, poor sanitation is a major public health challenge, contributing to 60% of diseases and 23% of under‐five mortality caused by diarrhea [7, 8, 9]. Despite efforts to improve sanitation, including the nationwide adoption of the Community‐Led Total Sanitation and Hygiene (CLTSH) program since 2011 which promotes latrine construction and hygiene behavior change significant gaps remain in latrine maintenance, water management, and sustained hygiene practices [10, 11, 12, 13]. Moreover, about 48 million Ethiopians still lack access to basic sanitation, underscoring the need for enhanced intervention strategies [14]. Inadequate sanitation also negatively affects social and economic development by compromising dignity, increasing the risk of gender‐based violence, reducing school attendance among girls due to poor menstrual hygiene management, and imposing substantial healthcare costs and productivity losses [13]. Furthermore, environmental consequences such as water pollution and biodiversity loss exacerbate public health risks and hinder sustainable development.

Despite these well‐documented impacts, Ethiopia continues to face low utilization of improved sanitation facilities with persistent regional and socioeconomic disparities. Many households still rely on unimproved or shared sanitation facilities, and open defecation remains prevalent in several areas [11]. Inconsistencies in the implementation of sanitation programs like CLTSH, especially regarding latrine maintenance and hygiene behavior sustainability, have limited the potential health benefits [15]. Additionally, existing studies often neglect the influence of community‐ and regional‐level factors, which can shape sanitation behaviors beyond individual household characteristics. This lack of comprehensive understanding hampers the design of effective, context‐specific interventions that can sustainably improve sanitation access and use across Ethiopia. Access to improved sanitation is thus not only a matter of infrastructure but also involves complex social, economic, and environmental determinants that operate at multiple levels.

Access to sanitation varies significantly across regions and socioeconomic groups in Ethiopia, reflecting complex, hierarchical influences at both household and community levels. Factors including education, wealth, household size, gender dynamics, and geographic location interact to shape sanitation behaviors and facility use [14]. However, many studies focus primarily on individual or household‐level factors without accounting for community or regional context, which may mask important structural and environmental influences on sanitation use. To address this gap, this study employs a multilevel logistic regression approach to investigate the prevalence and determinants of improved household sanitation use in Ethiopia, capturing both individual and community level variations. This methodology allows for a more nuanced understanding of sanitation behaviors and supports the development of tailored interventions to advance Sustainable Development Goal 6 universal access to improved sanitation by 2030.

2. Methodology

2.1. Study Area and Population

This study was conducted in Ethiopia, a landlocked country in the Horn of Africa, bordered by Eritrea, Djibouti, Somalia, Kenya, South Sudan, and Sudan. With an area of approximately 1.1 million square kilometers, Ethiopia has a population of about 132 million, making it the second most populous country in Africa after Nigeria. The capital, Addis Ababa, is located near the East African Rift Valley. Ethiopia is geographically and culturally diverse, and sanitation access varies considerably across regions. The study population included households that participated in the 2019 Ethiopian Mini Demographic and Health Survey (EMDHS), conducted between March 21 and June 28, 2019.

2.2. Source of Data

The data were obtained from the 2019 EMDHS, implemented by the Ethiopian Public Health Institute (EPHI) in collaboration with the Federal Ministry of Health (FMoH) and the Central Statistical Agency (CSA), with technical and financial assistance from ICF and other partners. The survey employed a stratified two‐stage cluster sampling design, ensuring national and regional representativeness. At the first stage, enumeration areas (EAs) were selected using probability proportional to size (PPS) within 21 strata (each region divided into urban and rural strata). At the second stage, a fixed number of households were randomly selected from each EA. This study included all 6141 households with complete data on sanitation facility use.

2.3. Study Design

This study used a cross‐sectional design based on data from the 2019 Ethiopian Mini Demographic and Health Survey (EMDHS). The survey employed a stratified two‐stage cluster sampling method using the 2019 Population and Housing Census as a frame. A total of 6141 households were selected from enumeration areas across all regions to provide nationally representative data on sanitation use.

2.4. Study Variables

The dependent variable in this study was the use of improved sanitation facilities, coded as a binary variable (1 = household uses improved sanitation; 0 = Household uses unimproved sanitation). Independent variables included a range of individual‐level factors such as marital status, sex, age, education, household wealth status, home ownership, and latrine sharing status. Community‐level variables comprised place of residence (urban vs. rural) and region.

2.5. Operational Definition

Improved sanitation service: refers to sanitation facilities that hygienically separate human excreta from human contact, such as flush toilets, ventilated improved pit latrines, and composting toilets.

Unimproved sanitation service: includes facilities that do not ensure safe separation, such as pit latrines without a slab, hanging latrines, or open defecation.

2.6. Data Extraction and Processing

A total of 6141 household heads were included after excluding households with missing data. Relevant variables were extracted and the data were exported from SPSS version 25 to Stata version 17 for analysis. Sampling weights were applied to adjust for varying selection probabilities across strata. Given the hierarchical nature of the 2019 EMDHS data set, with households nested within enumeration areas (EAs), the study treated EAs as level 2 and households as level 1 in the multilevel analysis.

2.7. Data Cleaning and Processing

Data cleaning and recoding were initially performed in SPSS version 25, followed by export to Stata version 17 for analysis. Sampling weights provided by the DHS were applied to adjust for the unequal probability of selection and ensure national representativeness. Households with missing values on key variables were excluded through complete‐case analysis to maintain data integrity. The hierarchical nature of the data with households nested within clusters was explicitly accounted for in the analysis.

2.8. Multilevel Logistic Regression Modeling

We used a two‐level multilevel logistic regression model to analyze the likelihood of improved sanitation use among households (level 1) nested within communities or clusters (level 2) [16]. A multilevel model was applied to the hierarchically structured data to simultaneously examine the effects of group level and individual level variation dependence on observations within and between groups. The model incorporated a random intercept to capture the variation between clusters, allowing us to account for the hierarchical structure of the data. The model equation can be expressed as: Logit (P_ij ) = β₀+ ${\sum }_{k=1}^k\beta$ _k X _kij + u_j ,

Where P_ij represents the probability that household i in cluster j uses improved sanitation, β₀ is the fixed intercept, β _k are fixed‐effect coefficients for predictor variables X_kij , and u_j is the random effect for cluster j, assumed to be normally distributed with mean zero and variance σ²u. This structure allows for the estimation of both within‐cluster (household‐level) and between‐cluster (community‐level) variation in sanitation use. The statistical model employed in this inquiry was the multilevel regression model [17].

2.9. Model Diagnostics

Before fitting the multilevel model, we conducted χ ² tests to evaluate heterogeneity across clusters. We also estimated the Intraclass Correlation Coefficient (ICC) from the null model to determine the proportion of total variance explained by between‐cluster differences. Likelihood ratio tests were used to compare the multilevel logistic regression model against a single‐level logistic regression, with the multilevel model demonstrating significantly better fit (p < 0.001). To assess potential multicollinearity among predictors, Variance Inflation Factors (VIFs) were calculated and all were below 2, indicating no substantial collinearity.

2.10. Estimation Method and Model Selection

Data analysis was conducted using Stata 17. Multilevel models were estimated using maximum likelihood estimation with adaptive quadrature to improve accuracy. Model selection was guided by information criteria, specifically Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), with the final model chosen based on the lowest AIC and BIC values, balancing model fit and parsimony.

3. Result

This study included a weighted sample of 6141 household heads in Ethiopia. The results show that male‐headed households are more likely to use improved sanitation facilities (80.9%) compared to female‐headed households (54.3%). Younger household heads, particularly those aged 20–30 years, are more likely to use improved sanitation (83.6%) and 51–60 years (69.4%), while the 31–40 age group shows a nearly equal split between improved and unimproved sanitation. Urban households have a notably higher rate of improved sanitation use (82.9%) than rural households (62.3%). This indicates that urban areas benefit from better infrastructure and services, while rural households lag behind. Households headed by individuals with higher education have the highest prevalence of improved sanitation (90.5%), followed by those with secondary (83.7%) and primary education (70.4%). Illiterate household heads have the lowest usage (53.5%), suggesting that education improves awareness and ability to access improved facilities. Wealth status further underscores disparities, with 79.3% of rich households using improved sanitation compared to only 54.6% among poor households. Home ownership shows a minor difference in improved sanitation use, with non‐owners slightly more likely (65.5%) than owners (63.9%) to have improved sanitation. Married households have the highest prevalence of improved sanitation use (70.5%), whereas unmarried (58.5%) and divorced/separated (55.4%) households have lower rates as presented in Table 1.

Table 1.

Socio‐demographic variables associated with use of improved sanitation services in Ethiopia.

Variable	Categories	Precise to use of sanitation service of household
		Improved		Unimproved		Total
		Count	Percent	Count	Percent	Count	Percent
Sex	Male	3760	80.9%	890	19.1%	4650	75.7%
	Female	810	54.3%	681	45.7%	1491	24.3%
Age	20–30	2059	83.6%	403	16.4%	2462	40.1%
	31–40	1258	50.1%	1254	49.9%	2512	40.9%
	41–50	27	58.7%	19	41.3%	46	0.7%
	51–60	737	69.4%	325	30.6%	1062	17.3%
	61–69	34	57.6	25	42.4	59	1%
Place of residence	Rural	2960	62.3%	1788	37.7%	4748	77.3%
	Urban	1155	82.9%	238	17.1%	1393	22.7%
Education level	Illiterate	1330	53.5%	1157	46.5%	2487	40.5%
	Primary	1666	70.4%	701	29.6%	2367	38.5%
	Secondary	564	83.7%	110	16.3%	674	11%
	Higher	555	90.5%	58	9.5%	613	10%
Wealth status	Poor	1499	54.6%	1247	45.4%	2746	44.7%
	Middle	866	73%	321	27.0%	1187	19.3%
	Rich	1750	79.3%	458	20.7%	2208	36%
Homeownership	Yes	2980	63.9%	1680	36.1%	4660	75.9
	No	970	65.5%	511	34.5%	1481	24.1%
Marital status	Unmarried	340	58.5%	241	41.5%	581	9.3%
	Married	2750	70.5%	1150	29.5%	3900	63.5%
	Divorce/separate	460	55.4%	370	44.6%	830	13.7%
	Widowed	550	67.1%	280	32.9%	820	13.4

Figure 1 shows the prevalence of improved and unimproved sanitation service use among households across different regions of Ethiopia. The result reveal significant regional variations, with Addis Ababa having the highest number of households using improved sanitation facilities, far exceeding those using unimproved sanitation. Similarly, Dire Dawa also shows a high prevalence of improved sanitation use compared to unimproved. In contrast, regions such as Somali and Afar have considerably lower numbers of households with improved sanitation, with unimproved sanitation use being more common in these areas. Other regions, including Tigray, SNNPS, Oromia, Amhara, Benishangul, Gambela, and Harari, show a higher number of households using improved sanitation than unimproved, although the difference is less pronounced compared to the capital and Dire Dawa. These findings highlight pronounced disparities between urban or more developed regions and predominantly rural or pastoral regions, indicating that households in urban centers have better access to improved sanitation facilities, while those in remote or less developed areas face challenges in obtaining adequate sanitation services.

Figure 1.

Illustrates the distribution of improved and unimproved sanitation service usage among households across various regions of Ethiopia.

3.1. Binary Logistic Regression Analysis

In the binary logistic regression analysis, each independent variable household heads' marital status, age, sex, education level, place of residence, home ownership, latrine sharing status, and wealth status was individually found to be significantly associated with sanitation use at the 5% significance level as presented in Table 2.

Table 2.

Results of univariate binary logistic regression analysis identifying potential factors associated with use of improved sanitation services, Ethiopia.

Variable	Log likelihood (χ 2)	df	Prob > χ 2
Region	−7871.9739	10	0.001
Residence	−8567.869	1	0.000
Educational level	−8347.876	3	0.001
Age	−7833.876	3	0.002
Wealth index	−8652.566	2	0.000
Marital status of household	−7896.765	3	0.130
Sex	−1278.578	1	0.014
Latrines not shared with other households	−4788.347	1	0.000
Homeownership	−2347.674	1	0.005

3.2. Test of Heterogeneity in the Use of Improved Sanitation Services Across Regions of Ethiopia

To account for unobserved heterogeneity arising from the hierarchical nature of the data specifically, households nested within regions a multilevel logistic regression model was employed. Before the multilevel analysis, a χ ² test was conducted to assess whether there was significant variation in the use of improved sanitation services across regions. As shown in Table 3, the χ ² statistic was 54.81 with a p‐value of 0.012, indicating statistically significant regional differences. This suggests that the likelihood of using improved sanitation facilities varies by region and that a single‐level model may not sufficiently capture this variability. Further evidence of regional clustering is provided by the intra‐class correlation coefficient (ICC), which was estimated at 0.035 with a p‐value of 0.015. This ICC value implies that approximately 3.5% of the total variation in improved sanitation use can be attributed to differences between regions. The random intercept variance (σ² = 0.196, p = 0.023) also supports the presence of between‐region variability. Together, these results justify the use of a multilevel logistic regression model to appropriately account for the clustering of households within regions and to obtain more accurate estimates of the effects of individual and community‐level predictors on sanitation use.

Table 3.

χ ² Test of heterogeneity in the use of improved sanitation services across regions of Ethiopia.

Improve Sanitation Service	β	Std. err	z	p
Fixed
${\beta }_0$‐intercept	−2.641	0.071	−33.80	0.000*
Random part
sigma(${{\delta }^2}_{\mathit{uo}}$) ‐ variance	0.196	0.086	2.27	0.023*
Intra‐region correlation coefficient
ICC (Rho (ρα))	0.035	0.010	4.81	0.015*
Likelihood‐ratio test of rho = 0 chiqar2(01) = 54.81 Prob >= chibar2 = 0.012*

Abbreviation: ICC, intra‐region correlation coefficient.

^*Significant at 5%.

3.3. Comparison Among Multilevel Logistic Regression Models

To determine the most appropriate statistical model for analyzing the factors associated with improved household sanitation use in Ethiopia, three multilevel logistic regression models were compared: an empty model with a random intercept (null model), a random intercept model with fixed effects, and a random coefficient model. Model fit was assessed using key statistical criteria, including the log‐likelihood, deviance, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). As presented in Table 4, the random intercept model with fixed effects demonstrated the best overall fit, with the lowest AIC (385.068) and BIC (491.294) values among all models considered. This indicates that incorporating fixed effects while accounting for regional clustering through a random intercept significantly improved model performance compared to the empty model and the more complex random coefficient model. The deviance χ ² statistic and corresponding p‐values further supported this conclusion, with the fixed‐effects model achieving statistical significance (p = 0.002). In contrast, the random intercept and random coefficient models yielded higher AIC and BIC values, indicating poorer fit. Therefore, the random intercept model with fixed effects was selected for final analysis, as it best captures both individual and regional level variations in the use of improved sanitation services across Ethiopia.

Table 4.

Model comparison for multilevel logistic regression models.

Methods	Random intercept	Random Intercept with fixed effect	Random coefficient model
Log Likelihood	−7993.647	−163.534	−2749.775
Deviance based on χ 2	4.813	1056.410	1.074
p‐value	0.345	0.002*	0.167
AIC	15991.300	385.068	5503.551
BIC	16006.460	491.294	5516.561

Abbreviations: AIC = Akaike Information Criteria, BIC = Bayesian information criteria.

^*Significant at 5% level.

Table 5 presents the results of the multilevel logistic regression model, incorporating both fixed effects (individual‐level covariates) and random effects (regional‐level variance). The inclusion of level‐one covariates such as education, wealth, marital status, and place of residence reduced the regional‐level variance from 0.473 in the intercept only model to 0.265 in the final model. This reduction indicates that a significant portion of the variation in improved sanitation use across regions can be explained by differences in household‐level characteristics. The decrease in random intercept variance demonstrates that adding fixed explanatory variables improves the model's predictive ability and confirms the presence of substantial regional disparities in sanitation access.

Table 5.

Summary of random intercept with fixed slope multilevel logistic regression model analysis.

Parameter estimate	Coefficient	SE	Z	p‐value	Odds ratio
Fixed effect intercept(βo)	−0.413	1.868	−0.3	0.022*	0.715
Random effect var(UAnj)	0.2650	0.391
Intercept only model var(region)	0.473	0.731

Abbreviation: SE = Standard error.

^*Significant at 5% levels.

After adjusting for regional heterogeneity, several variables emerged as statistically significant predictors of improved sanitation use at the 5% significance level. These include household wealth status, education level, marital status, place of residence (urban vs. rural), home ownership, and whether the household uses shared or private latrines. Each of these factors independently contributes to the likelihood of a household accessing improved sanitation services. In contrast, the age and sex of the household head did not show a statistically significant association with sanitation use in the final model. This suggests that socioeconomic and environmental factors play a more critical role than demographic factors in influencing sanitation access in Ethiopia.

3.4. Interpretation of Multilevel Logistic Regression Results

The multilevel logistic regression analysis presented in Table 6, using a random intercept model, identifies key socio‐demographic and household‐level factors significantly associated with the use of improved sanitation services among Ethiopian households, while accounting for regional clustering effects. Marital status emerged as a strong predictor of improved sanitation use. Compared to widowed household heads (the reference group), married individuals were significantly more likely to use improved sanitation facilities, with 1.81 times higher odds (OR = 1.810; 95% CI: 1.67–4.698; p < 0.001). In contrast, households headed by unmarried individuals had 65% lower odds (OR = 0.350; p < 0.001), and those headed by divorced individuals had 76% lower odds (OR = 0.240; p = 0.002), indicating that being married may provide greater economic stability or household support, which enhances access to sanitation.

Table 6.

Factors associated with improved sanitation use among Ethiopian households: multilevel logistic regression results (random intercept model).

Variable	Categories	Odds ratio (AOR)	SE	p‐value	95% CI for OR
Marital status	Widowed (Ref)	1	—	—	—	—
	Married	1.810	0.264	—	—	—
	Unmarried	0.350	0.249	0.000*	0.212	0.563
	Divorce	0.240	0.237	0.002*	0.210	0.498
Residence	Urban (Ref)	—	—	—	—	—
	Rural	0.370	0.193	0.043	0.140	0.996
Educational level	Higher education (Ref)	—	—	—	—	—
	Illiterate	0.125	0.227	0.000*	0.080	0.196
	Primary	0.251	0.221	0.000*	0.162	0.387
	Secondary	0.573	0.255	0.159	0.348	1.944
Wealth index	Rich (Ref)	—	—	—	—	—
	Poor	0.860	0.143	0.009*	0.440	0.976
	Middle	0.841	0.145	0.016*	0.424	0.934
Homeownership	No (Ref)	—	—	—	—	—
	Yes	2.280	0.216	0.001*	1.126	2.456
Latrines not shared with other households	No (Ref)	—	—	—	—	—
	Yes	1.430	0.450	0.005*	1.361	2.651

Note: Ref = indicate reference and (Clustering effect or regional heterogeneity considered).

^*Significant at 5% level.

Place of residence was another significant factor. Households in rural areas had 63% lower odds of using improved sanitation compared to their urban counterparts (OR = 0.370; 95% CI: 0.94–0.996; p = 0.043). This finding reflects persistent urban‐rural disparities in infrastructure availability and access to public health services. Education level showed a strong and statistically significant association. When compared to household heads with higher education (reference group), those who were illiterate had 87.5% lower odds of using improved sanitation (OR = 0.125; p < 0.001), and those with only primary education had 74.9% lower odds (OR = 0.251; p < 0.001).

Wealth status was another critical determinant. Households in the poor and middle‐income categories had significantly lower odds of using improved sanitation compared to those in the rich households. Poor households had 14% lower odds (OR = 0.860; p = 0.009), while middle‐income households had 16% lower odds (OR = 0.841; p = 0.016). Home ownership was positively associated with improved sanitation use. Households that owned their homes had more than twice the odds of using improved sanitation services compared to non‐owners (OR = 2.280; p = 0.001). This may reflect greater ability or autonomy to invest in durable sanitation facilities when ownership is secured. Furthermore, households with private latrines (not shared with other households) had significantly higher odds of improved sanitation use (OR = 1.43; p = 0.0048), reinforcing the idea that shared facilities are often inadequate or unimproved.

4. Discussion

This study examined the prevalence and determinants of improved sanitation use among Ethiopian households using data from the 2019 Ethiopian Mini Demographic and Health Survey and a multilevel logistic regression approach to account for regional variations. Significant regional variation in sanitation use was confirmed by the χ ² test and the intra‐class correlation coefficient (ICC), which justified the application of a multilevel modeling framework. This regional heterogeneity reflects differences in infrastructure availability, local governance, resource allocation, and possibly cultural practices across Ethiopia's diverse regions. Addressing these regional disparities is essential for equitable progress toward national sanitation goals.

Marital status emerged as a significant determinant of improved sanitation use, with married households being more likely to have access compared to widowed, unmarried, or divorced households. This finding is consistent with studies from Ethiopia and broader sub‐Saharan Africa, including those by [18, 19, 20, 21], which link higher sanitation access among married individuals to increased economic stability, shared household responsibilities, and a greater focus on family well‐being. Married couples often have stronger financial capacity and motivation to invest in sanitation infrastructure, whereas widowed or single‐headed households may encounter economic hardships, social isolation, or limited autonomy in decision‐making, all of which can impede access to improved sanitation facilities. These findings highlight the need for targeted sanitation interventions that consider household structure, particularly to support vulnerable groups such as widowed or single‐headed households, to address disparities in sanitation access and promote equitable health outcomes.

The educational level of the household head significantly influences the use of improved sanitation services. Households with lower levels of education are less likely to utilize improved sanitation facilities compared to those with higher education. This finding aligns with previous research indicating that households with higher educational attainment are more likely to adopt improved sanitation practices [22, 23]. Educational attainment is a strong predictor of sanitation use, as it enhances health knowledge and empowers households to prioritize sanitation. Notably, the largest gap in sanitation use is observed between individuals with higher education and those who are illiterate, underscoring education's critical role in sanitation behavior. Education likely increases awareness of hygiene and the health benefits associated with improved sanitation, encouraging households to invest in and maintain better facilities. Furthermore, the study revealed that nearly one‐third of the surveyed population lacked formal education, consistent with findings [24], who reported a strong positive association between education level and sanitation facility usage. Low education levels in many households may contribute to poor sanitation behaviors and lower adoption of improved services. These findings emphasize the importance of education in promoting public health interventions and highlight the need for targeted educational programs to improve sanitation practices, particularly among less educated populations.

Rural households were significantly less likely to use improved sanitation facilities compared to urban households. This rural‐urban gap highlights the persistent challenges in extending sanitation infrastructure and services to remote and underserved rural areas, where access to clean water and health education is often limited. The rural‐urban disparity observed in this study aligns with findings from multiple previous studies, including [25, 26], which reported that sanitation services are predominantly concentrated in urban areas due to better infrastructure, higher incomes, and greater public awareness. These studies also emphasize that rural households face substantial logistical and financial barriers that hinder improvements in sanitation facilities.

Wealth status is a critical determinant of improved sanitation use, reflecting the broader socioeconomic dynamics that influence health‐related behaviors and access to infrastructure. Households with greater economic resources are better positioned to invest in the necessary materials and facilities required for improved sanitation, such as constructing private latrines and ensuring their proper maintenance. This economic advantage reduces barriers related to upfront costs and ongoing upkeep, which often prevent poorer households from adopting or sustaining improved sanitation solutions. The findings align with previous studies, including those by [27, 28], which similarly highlight how financial capacity directly impacts the ability to secure and maintain sanitation facilities. Moreover, wealthier households may also benefit from better access to information, markets, and services, further facilitating their sanitation improvements. This underscores the need for targeted interventions that not only provide infrastructure but also address economic inequalities, such as subsidies, microfinance, or community‐based financing schemes, to enable low‐income households to access and maintain improved sanitation.

Home ownership significantly influences the use of improved sanitation facilities among Ethiopian households. Those who own their homes are more likely to utilize improved sanitation services compared to renters or those living in leased housing. This aligns with evidence indicating that families in privately owned homes are nearly three times more likely to have access to improved sanitation [29]. The role of home ownership as a facilitator of sanitation investment is further supported by [30], who linked tenure security with improved household infrastructure. The lower prevalence of improved sanitation among renters may be due to their generally lower socioeconomic status, which limits financial capacity to invest in sanitation upgrades. Moreover, the often poor condition of rental properties restricts tenants' ability or willingness to make such improvements. This socioeconomic vulnerability presents a significant barrier to sanitation access. Addressing these disparities requires targeted interventions such as subsidized sanitation facilities or financial assistance programs to enable low‐income and renting households to access and maintain improved sanitation services [31]. Such measures are essential for promoting equitable sanitation access and improving public health outcomes across all housing types in Ethiopia.

The study found that households with private latrines are significantly more likely to use improved sanitation facilities, aligning with research by [1, 32], which show shared latrines are often less hygienic and less used. Private latrines offer households greater control over hygiene and maintenance, which are essential for sustained use and maximizing health benefits. Shared sanitation facilities, while often seen as a practical solution in resource‐limited settings, tend to suffer from neglect, overcrowding, and poor cleanliness, which can discourage consistent use and increase the risk of disease transmission. This is particularly important for vulnerable groups such as women and children, who may experience safety and privacy concerns when using shared facilities, leading to avoidance behaviors that increase exposure to unsafe sanitation practices like open defecation. Therefore, sanitation policies must prioritize access to private or small‐group latrines to address these challenges effectively. Moreover, improving the quality and management of shared facilities where private options are not feasible can help mitigate some of these issues. Ultimately, focusing on private sanitation access not only promotes better hygiene but also supports dignity and safety, contributing to broader public health and social equity goals.

4.1. Strengths and Limitations of the Study

This study's strength lies in using nationally representative data, allowing for generalizable findings across diverse regions of Ethiopia. The use of multilevel logistic regression effectively accounted for regional heterogeneity and clustering of households within regions, providing more accurate estimates of factors influencing improved sanitation use. Additionally, incorporating both individual and household‐level variables enriched the analysis, offering a comprehensive understanding of the socio‐demographic determinants of sanitation access. However, its cross‐sectional design limits causal inference, and some important factors like hygiene practices and local infrastructure were not included. Additionally, reliance on self‐reported data may introduce bias.

5. Conclusion

This study reveals substantial regional disparities in the use of improved sanitation services among Ethiopian households, highlighting the need to consider geographic heterogeneity in the design and implementation of sanitation interventions. Key factors significantly associated with improved sanitation use include marital status, residence (urban vs. rural), education level, household wealth, home ownership, and whether the household uses a private or shared latrine. Households that are married, urban‐based, better educated, wealthier, and own their homes are more likely to access improved sanitation. In contrast, widowed, rural, less educated, poorer, and non‐homeowner households face greater barriers. These findings underscore the complex, multidimensional nature of sanitation access and the need to address intersecting social, economic, and infrastructural inequalities.

To enhance sanitation coverage and promote equity, comprehensive interventions are required. These should focus on expanding education, fostering economic empowerment, and investing in infrastructure particularly for vulnerable groups such as widowed, single, and low‐income households. Promoting private latrine ownership and improving the safety and quality of shared facilities is also essential, especially given concerns about privacy and security, particularly among women.

Author Contributions

Dagne Tesfaye Mengistie contributed to the design of the analysis, the thorough writing of the article, the critical drafting for significant intellectual interaction and contributed to designing the data analysis plan and interpreting the findings. Teshome Bekele Elama contributed to data collection and played a key role in conceptualizing the research questions and study objectives. Buzuneh Tasfa Marine contributed to the conceptualization and design of the study, oversaw the overall design of the data analysis, provided substantial input in drafting and revising the manuscript, and critically evaluated the manuscript for significant intellectual contributions ensuring clarity, accuracy and scientific rigor. Buzuneh Tasfa Marine also coordinated the final approval process and managed the submission of the manuscript. All authors have read and approved the final version of the manuscript. Buzuneh Tasfa Marine accepts full responsibility as the guarantor for the integrity of the data and the accuracy of the data analysis presented in this study.

Ethics Statement

This study was carried out in accordance with the relevant guidelines and regulations outlined in the Helsinki Declaration. The Research Ethical Review Committee of Jigjiga University provided ethical approval. The informed consent was obtained from study participant. They were guaranteed the right to refuse or withdraw from the study whenever they desired.

Consent

No individual person's personal details, images, or videos are being used in this study.

Conflicts of Interest

The authors declare no conflicts of interest.

Transparency Statement

The lead author Buzuneh Tasfa Marine affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Mengistie D. T., Elama T. B., and Marine B. T., “Determinants of Improved Household Sanitation Use in Ethiopia: A Multilevel Logistic Regression Analysis,” Health Science Reports 8 (2025): 1‐10, 10.1002/hsr2.71544.

Data Availability Statement

The data sets used and or analyzed during the current study are available from the corresponding author on reasonable request.