Quality of latrines and willingness to improve them in rural Ethiopia

IBM-WASH

integrated behavioural model for water, sanitation, and hygiene

OD

open defecation

RANAS

risks, attitudes, norms, abilities, and self-regulation

SD

standard deviation

With the most significant decrease in the estimated national open defecation (OD) rate worldwide from 79 to 22% between 2000 and 2017, Ethiopia has been praised for its remarkable progress in improving sanitation conditions (UNICEF/WHO 2019). Less attention has nevertheless been given to the unsettling fact that the sanitation change has largely comprised a shift towards the utilization of unhygienic latrines. According to the 2017 national estimates, 82% of sanitation facilities were unimproved (did not safely prevent excreta from human contact) and only 7% of latrines in rural Ethiopia were considered as safely managed latrines (UNICEF/WHO 2019). The generally low quality of latrines in Ethiopia was also documented by several field level cases studies (Awoke & Muche 2013; Irish et al. 2013; Crocker et al. 2017; Novotný et al. 2017, 2018a; Zeleke et al. 2019). A similar pattern is prevalent across some other countries of Sub-Saharan Africa and represents a major reason why implemented sanitation interventions often fail to prevent the transmission of pathogens (Irish et al. 2013; Exley et al. 2015; Sclar et al. 2016).

The increased latrine coverage in Ethiopia can be attributed to a national hygiene and sanitation strategy (MoH 2013) implemented through the country-wide health extension programme (MoH 2015; Assefa et al. 2019). A key component was the community-led total sanitation approach which is known to be effective for rapidly eliminating OD but problematic regarding the quality and durability of adopted sanitation facilities (Venkataramanan et al. 2018; Ficek & Novotný 2019). Facilitation of access to sanitation hardware had been originally included in the national Ethiopian strategy, but it was side-lined in practice (Novotný et al. 2018a). Low quality and undurability of latrines not only challenge their presumed positive impacts on human health but also contribute to the slippage back to OD (Crocker et al. 2017; Abebe & Tucho 2020; Aragie et al. 2022; Freeman et al. 2022).

Only a very few studies examined issues around latrine upgrading in Ethiopia (Novotný & Mamo 2022) and they report no or little shifts upwards through the sanitation ladder (Crocker et al. 2017; Chambers et al. 2021; Mamo et al. 2023). Resources invested in sanitation interventions may be wasted if basic hygienic standards of latrines are not ensured. Understanding the behaviours, preferences, and priorities of households with respect to latrine quality and its improvements is thus an important task. This task is addressed in this study based on data from the Loka Abaya district, Sidama region, through direct observations of latrines and structured interviews among 422 latrine-owning households. The aim is to examine latrine quality and latrine upgrading and the respective plans and preferences.

The interview schedule consisted of 57 questions covering demographic and socioeconomic characteristics; questions on latrine adoption, use, maintenance, past improvements, and plans to improve in the near future; sanitation-related preferences, availability of sanitation-related services; and psychosocial measures related to sanitation. Direct observations of sanitation facilities and their surroundings assessed 19 parameters.

All participants and informants participated voluntarily in the survey based on their oral consent sought after an introductory description of the survey and its purpose. Participants were assured of their anonymity and the confidentiality of the gathered information. Ethical clearances were obtained from the Ethical Committee of the Wolaita Sodo University and the Ethical Committee of Charles University. Approval was also obtained from the district authorities.

The descriptive findings represent an important part of this study. In addition, we analysed the correlates of latrine quality and willingness to improve latrines using linear regression and binary logistic regression, respectively. The analysis was conducted using a Complex Samples module in SPSS, while accounting for the sampling frame (clustering of data). Latrine quality was measured by the aggregate latrine quality score constructed based on 11 distinct parameters of latrine quality measured through the direct observations of latrines and their surroundings. We explored the dimensionality of the set of these parameters and identified six underlying dimensions that were eventually used for the construction of the composite score (see Supplementary materials S1). Although the latrine quality score is a bounded outcome that can attain values between zero and six, we still used liner regression for modelling this outcome after checking that the distribution of the regression residuals is approximately normal and that the fitted values are not extrapolated outside of its range (see Supplementary materials S2).

The willingness to improve was elicited by a question, ‘Do you plan to improve your latrine in the near future?’ measured on a five-point Likert scale. We could not use ordinal regression to examine its predictors as the proportional odds assumption was not met. We thus used the binary logistic regression by considering a dichotomous outcome that distinguished between those who reported intention to improve their latrines against those who didn't.

We conceptually distinguish the following three subgroups of explanatory variables: (1) general objective characteristics of respondents and their households (demographic, socioeconomic, sociocultural, and ecological), (2) sanitation-related objective characteristics (latrine quality variables, characteristics of the process of adoption, maintenance, and improvement of latrines based on retrospective recalls), and (3) sanitation-related psychosocial variables (knowledge, perceived health risks, satisfaction, attitudes and preferences, descriptive and injunctive social norms, social identity, social cohesion). The IBM-WASH model categorises contextual, technology, and psychosocial factors (Dreibelbis et al. 2013). In addition, the RANAS model (Mosler 2012) provided a vital inspiration for the specification of psychosocial variables considered in this study. The description of the considered variables appears in Supplementary materials S3, together with their descriptive statistics. The directions of their expected relationships with the analysed sanitation outcomes are mostly intuitive. The exceptions are some demographic variables for which documented effects can lead in both directions due to multiple possible underlying mechanisms (Novotný et al. 2018b, p. 129).

Specification of regression models was not a straightforward task as we worked with a moderate sample size and a larger number of potentially relevant predictors of which some were correlated. We could not include them altogether in a single model. We proceeded in the following steps. In the first step, we considered demographic and socioeconomic variables only. In the second step, we excluded some of the variables that were insignificant in the first step and included the three sanitation-related objective variables. In the third step, we excluded some other more insignificant variables and additionally included psychosocial variables (those for which multicollinearity was acceptable).

Of 549 households that were originally sampled, 422 owned latrines, which means 77% latrine coverage. The 62% of respondents from latrine-owning households reported that they always use their latrine for defecation when at home or nearby, and another 22% stated that they mostly use their latrine. The remaining 16% admitted that they use latrines only sometimes or rarely in at least one of the time spells that we asked for (dry and rainy seasons and day-time and night-time). The latrine quality score for this subgroup was significantly lower than for the rest of the respondents. The findings above imply that around 38% of households in the study area used OD as their main defecation practice. This is not far from the respondents’ estimates of the OD rates in their villages – on average, they estimated an OD rate of 30%. In addition, latrine use was considerably lower for children (aged 5–15), particularly in the rain-season and night-time. More than half of latrine-owning households with children admitted that their kids rarely use latrines.

The supply of hygienic latrine components (slabs in particular) and sanitation services was almost non-existent in the study area. Only a minor share of 14% of households reported that they have ever used some sanitation services. These predominantly involved general tasks such as the digging of pits or crafting or fixing latrine superstructures.

Table 1 presents results from the analysis of correlates of latrine quality score. Two demographic variables in terms of family size and the presence of children under 5 years of age were identified as the statistically significant predictors with positive relationships to latrine quality in the first and second (for family size) models. However, their regression coefficients were relatively weak and became statistically insignificant when we additionally included psychosocial variables in the third regression model. Of the socioeconomic variables, only those that distinguished households that primarily depend on casual labour (agricultural wage labourers) from the rest of households (mostly farmers) revealed a negative relationship with latrine quality. The effect was comparatively strong and statistically significant across all three model specifications. Expectably, two variables that recorded whether there had been any improvements in latrines in the past 2 years were positively correlated with the latrine quality score. This was not the case for the reported use of sanitation services, which may be explained by the nature of these services (digging the pits and/or basic construction works).

The sanitation-related psychosocial variables were comparatively more important predictors of latrine quality (model 3 in Table 1). Interestingly, the measure of perceived vulnerability to diarrheal disease showed a negative relationship with latrine quality, while that for perceived severity was positive. The variables of injunctive but not descriptive social norms were statistically significant in terms of norm perception. Satisfaction with the current sanitation practices was positively related to latrine quality. On the other hand, some intuitively relevant psychosocial variables, such as measures of hygiene and sanitation knowledge, recognition of the benefits of latrines (their various types), or social context variables (social trust, identification, cohesion), were not statistically significant predictors of latrine quality.

Despite the generally poor quality of latrines, 63% of respondents reported satisfaction with their current sanitation practice, and nearly a quarter of respondents didn't recognize (reported) any disadvantages or weaknesses of their latrines. The most frequent problems and disadvantages were the nondurability of latrine (32% of respondents), flooding (26%), bad smell, concentration of flies, and maintenance costs (each mentioned by 25–30% of respondents). By contrast, concerns about possible safety risks, health risks, or adverse effects of latrines on the environment were considerably less frequent (mentioned by less than 5% of respondents). These observations can be directly related to the fact that the majority of respondents (61%) explicitly stated that they believe that the use of their latrine helps them to avoid diseases.

In addition, we asked respondents to compare the importance of particular components of pit latrines prevalent in the study area (Table 2). Pit size was clearly perceived as the most important latrine attribute, followed by solid walls, a solid slab platform, and functional doors.

Only 10% of households in the sample adopted latrines recently, while for 43% it was more than 1 year ago, and for 40% it was more than 5 years before our survey. For nearly all households in our sample (94%) the present latrine was not their first latrine. The usual durability of sanitation facilities in the surveyed area was between 2 and 5 years. When a pit gets full, it is covered by soil and a new pit is dug elsewhere.

Nearly two-thirds of households (70%) stated that their current latrines are better than their initial sanitation facilities. Although this might indicate possible gradual upgrading, qualitative descriptions of recent improvements suggested that these are mostly limited to regular maintenance tasks. Table 3 classifies households according to the types of reported improvements. We can see that only less than a third of households didn't undertake any improvements and that this subgroup of households had significantly lower quality of their latrines. Nearly a third of households undertook the significant reconstruction of their latrines, which mostly comprised new latrine constructions either due to exhausting pit capacity or due to their collapses. The rest of the households (38%) reported various minor improvements. These were mostly repairs of walls, roofs, doors, or floors and less often also pits enlargements or solidifications. Interestingly, the average quality score for this subgroup was higher than for those who substantially reconstructed their facilities, though the difference was not statistically significant. It suggests that newly rebuilt latrines tend not to be of better quality than previous ones.

Nearly two-thirds of respondents (61%) reported an intention to improve their latrines in the near future (Table 4). Of 232 respondents who answered an open-ended question on the nature of planned improvements, 47% wanted to construct an entirely new latrine, 40% planned to improve the roof, 28% doors, 17% slabs, and 16% walls. Other specific plans were rarely reported.

Ensuring more privacy was the most frequent motivation for latrine improvement (expressed by 42% of those with an intention to improve latrine) followed by safety (36%), comfort (21%), and reduction of health risk (17%). On the other hand, the most frequent reason for the lack of willingness to improve was that the current latrine is satisfactory (reported by 41% of those with no intention to improve), followed by high costs (21%), and a lack of materials (12%). The relatively rare occurrence of health-related motivations for latrine improvement (17%) can be compared to the considerably more frequently reported health-related motivations for the initial latrine adoption (79%) and for the current latrine use (61%).

Table 4 shows that the predictive power of variables examined in this survey with respect to the willingness to improve latrines was generally low. None of the demographic and socioeconomic variables was statistically significant. There was a positive relationship between the willingness to improve and latrine quality identified in model 2, though it became statistically insignificant when we additionally included psychosocial variables in model 3. This is a counterintuitive finding which contradicts an expectation that people who have comparatively worse sanitation facilities would be more interested in their improvement. The perception of vulnerability to diarrhoeal disease was positively related to the willingness to improve the latrine. By contrast, knowledge of sanitation and hygiene messages was negatively associated with this outcome. Rather surprisingly, none of the other variables that were also examined as potentially consequential predictors of the willingness to improve latrines were found to be statistically significant.

The latrine coverage in our study area was 77%, which is close to 79% estimated for Ethiopia by the UNICEF/WHO Joint Monitoring Programme (UNICEF/WHO 2019). In our survey, 12% of latrine-owning households admitted OD as their main defecation practice. This can be compared to the 16% rate of OD-free slippage identified in a systematic review by Abebe & Tucho (2020). It means that basic sanitation indices revealed for our study area are similar to the national estimates and that the findings of this paper may have a wider relevance beyond the context of this study.

Previous research on household-level sanitation in Ethiopia predominantly examined the initial adoption and availability of latrines. In spite of the fact that the majority of sanitation facilities across rural Ethiopia are considered to be unsafe (UNICEF/WHO 2019), issues around the sustainability and upgrading of latrines have been much less studied (Novotný & Mamo 2022). Our survey also documented the generally low quality of inspected latrines. Nearly all sanitation facilities in the study area were unventilated pit latrines constructed from local and mostly nondurable materials. On average, they met only 2.8 of the total six latrine quality dimensions determined based on multiple latrine parameters observed in the survey.

Almost all households in our sample had already rebuilt their latrines at least once prior to our survey. More than two-thirds of them (70%) reported some improvements in their latrines in the past 2 years. For one-third of households, these changes mostly involved basic repairs of latrines’ superstructure, while another third rebuilt their latrines completely due to either collapses or full pits. The mean quality scores for both these subgroups were comparatively higher than for the rest of the households, but still quite low. Our results thus did not show any substantial shifts up through the sanitation ladder. Similar findings were documented by Crocker et al. (2017) or Mamo et al. (2023).

Despite the generally low quality of latrines in our sample, the majority of their users (63%) expressed satisfaction with their sanitation situation. Unlike in Novotný et al. (2017), satisfaction was positively related to latrine quality, indicating that users do acknowledge the benefits associated with comparatively better facilities. However, this recognition does not translate into their plans to improve latrines as suggested by the absence of a relationship between the willingness to improve and satisfaction with the current sanitation situation. In other words, an intuitive assumption that dissatisfaction with one's own sanitation situation would catalyse the interest in improving this situation seems to not hold in the context of our study. This inference might appear incompatible with the observation that the majority of households in our sample (61%) reported plans to improve their latrines in the near future. However, the willingness to improve was positively related to latrine quality, which indicates that the poor latrine quality is itself not a major motivation for its improvement. The reported plans to improve mostly referred to the reconstructions and maintenance tasks rather than to upgrade the functionality.

Accordingly, a large pit was clearly seen as the most important latrine attribute. As in Goddard et al. (2018), we found that the solid slab, which is commonly regarded as a key aspect of a hygienic pit latrine, was perceived as comparatively less important. Similarly, issues around the limited lifespan and durability, together with privacy for users, prevailed in the discussions of motivations for latrine improvement, while motivations to reduce health risks or enhance user safety were rarely mentioned. Interestingly, however, our respondents were generally convinced about the health benefits of their latrines. All these findings suggest that there is an inadequate understanding of how a hygienic latrine works for preventing the transmission of pathogens. Such a conclusion is further supported by the fact that the measures of hygiene and sanitation knowledge were not related to latrine quality in our survey and, moreover, their relationship with the willingness to improve was negative. There is an apparent need to enhance the understanding about what are the key attributes of hygienic latrines. This was underplayed in the previous sanitation campaigns in Ethiopia. It seems that they created a widespread but superficial normative perception of the inevitably positive health benefits of latrines, irrespective of their technical standards.

Research literature from other countries than Ethiopia show that the socioeconomic situation of households represents a major factor for the investments of households into hygienic sanitation (Gross & Günther 2014; Simiyu 2017; Turrén-Cruz et al. 2020; Tiwari et al. 2022). In our study, however, we did not identify any relationship between the latrine quality and household income, wealth or attained education. We think that it is related to the type of latrines used in the study area. Their construction requires some manpower, suitable space, and basic local materials. They were predominantly self-constructed and only very rarely contained commercially purchased components. For adopting these facilities, the socioeconomic situation of households does not appear to be a crucial barrier. A sole socioeconomic variable that was the statistically significant predictor of latrine quality in our regression analysis was the dependence of households on casual agricultural labour. Although reporting comparatively higher income, this subgroup of families revealed significantly lower quality of latrines. This may just be due to their worse access to land and local materials.

That socioeconomic factors did not predict the quality of latrines in our sample obviously does not mean that they are unimportant in the present context. The contrary is true with respect to the prospective upgrading of local sanitation conditions. The documented socioeconomic situation in the study area predetermines the generally very low purchasing power of the local population. It undermines the development of a local supply of hygienic sanitation products and services on a commercial basis. Such services were non-existent in the surveyed area at the time of our survey. Identifying and supporting local-level sanitation actors is obviously another key challenge. However, it is unlikely that this will be possible without bolstering the purchasing power of local people. While the no-subsidy approach was used to initiate latrine adoption in Ethiopia, financial and/or material incentives would be instrumental and very probably required for facilitating latrine upgrading (Gebremariam & Tsehaye 2019; Tamene & Afework 2021; Afework et al. 2022; Mamo et al. 2023).

Finally, let us mention some limitations of this study. First, we used cross-sectional data and the relationships identified in our regression analyses represent statistical associations only. Reverse causality may be particularly an issue when interpreting results on the role of psychosocial predictors. Second, the quality of latrines was measured through direct observations of latrine attributes. Although practically feasible, the latrine quality score considered here is obviously only a very crude proxy measure of the effectiveness of latrines regarding the reduction of pathogens as well as concerning the comfort and safety of users. Third, previous community-based sanitation campaigns implemented in the study area utilized social and institutional surveillance. It established a strong perception of the injunctive social norms around latrine use. As such, what the respondents reported on their sanitation-related behaviours and preferences may thus involve a social desirability bias.

This study examined the quality of latrines and their improvements at the household level in the Loka Abaya district, South Ethiopia. The former was analysed based on direct observations and the latter by retrospective recalls and questions on preferences and future plans concerning latrine upgrading. The quality of latrines in our sample was low, questioning their positive effects on human health. In spite of this fact, the majority of respondents reported satisfaction with their sanitation situation. The prevalence of past latrine improvements and the willingness to improve latrines in the near future were high. However, they rarely addressed substantial functional upgrading and mostly involved regular maintenance and reconstruction due to the low durability of sanitation infrastructure. Our findings showed that there is an inadequate understanding about what are the key attributes of hygienic latrines. Together with the low purchasing power of local people, it impairs demand for hygienic sanitation and prevents the development of the respective supply chains. It implies a need for sanitation subsidies or material support, community education about the means of hygienic sanitation, and promotion of hygienic sanitation infrastructure. Both sanitation interventions and research should concentrate more on the sustainability and quality of sanitation facilities and not solely on increasing latrine coverage.

We thank research participants for their willingness to take part in the study and for their time spent with us during the households’ visits. We gratefully acknowledge the enumerators for their effort and collaboration during data collection. We would also like to appreciate the Loka Abaya district health office for the consent to conduct the study.

B.G.M. designed the study, organized and supervised data collection, processed and analysed the data, and participated in the manuscript preparation. J.N. designed the study, analysed the data, and participated in the manuscript preparation. A.A. designed the study and participated in the manuscript preparation. All authors read and approved the final manuscript.

The authors gratefully acknowledge support from the Czech Science Foundation [Grant number GA19-10396S]. This support was particularly important in the stage of data collection. The funding body played no role with respect to the study design, collection, analysis, and interpretation of data, or in writing the manuscript.

All participants and informants participated voluntarily in the survey based on their oral consent sought after an introductory description of the survey and its purpose. Participants were assured of their anonymity and the confidentiality of the information gathered. Ethical clearances were obtained from the Ethical Committee of the Wolaita Sodo University and the Ethical Committee of the Charles University. A written approval was also obtained from the district authorities.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict. Although the first author was involved in sanitation projects implemented in some of the surveyed kebeles, the authors feel and declare that it does not impact the interpretation of findings in this article.