Determinants of diarrhoeal infections among users of shared sanitation in informal settlements in Durban, South Africa

Globally, urban populations are increasing at alarming rates, mainly in cities in low- and middle-income countries in Africa and Asia (Reymond et al. 2016). It has been projected that by 2100, the global population will be about 11.2 billion, with an estimated 39% in Africa (Reymond et al. 2016). The provision of sanitation has not kept up with the increasing urban population, especially in urban slums or informal settlements (Simiyu et al. 2017; WHO/UNICEF 2017). Even when the sanitation infrastructure is available, it covers only a small fraction of the urban population. The World Bank estimates that over 51% of the urban population in Africa relies on unimproved sanitation, 15% rely on improved latrines, and only 25% have access to connected sewers or septic tanks (Reymond et al. 2016).

The lack of access to sanitation has led to increased use and need for the sharing of sanitation facilities. It is estimated that globally over 761 million people use shared sanitation or public sanitation facilities (Fuller et al. 2014), constituting about 32% of global sanitation access (WHO/UNICEF 2014; Rheinländer et al. 2015). Shared sanitation has been encouraged especially in densely populated areas in urban settings (Schouten & Mathenge 2010; Katukiza et al. 2012). The use of shared sanitation is therefore expected to increase further due to rapid urbanisation (Moore et al. 2003). Africa has the highest rate of shared sanitation use (44.6%) (Heijnen et al. 2014). It is estimated that between 2.9 and 3.6 million people in South Africa live in informal settlements, representing close to 13.9% of the total population (Selebalo & Webster 2017). These households rely on shared sanitation or in extreme instances of open defaecation.

Reviews on the impact of shared sanitation on diarrhoeal infections have shown that the use of these facilities increases the risks of infections, compared to private sanitation facilities (Heijnen et al. 2014; Ramlal et al. 2019). Additionally, the lack of adequate sanitation facilities has been implicated by several researchers as the main factor behind diarrhoeal diseases (Cameron et al. 2013; Prüss-Üstün et al. 2014; Freeman et al. 2017), helminth infections, and trachoma (Guerrant et al. 2013; Strunz et al. 2014; Sclar et al. 2016). Diarrhoea accounts for the largest share of morbidity and mortality attributed to sanitation, responsible for over 1.4 million deaths annually (GBD 2018), with over 533,000 of these deaths occurring in children under 5 years of age (Troeger et al. 2020). Poor sanitation also has adverse effects on the nutritional status of young children, through the impaired absorption of nutrients (Humphrey 2009; Guerrant et al. 2013). Sanitation interventions alone are responsible for over 32% reduction in diarrhoeal incidence (Fewtrell et al. 2005), and access to sanitation lowering the odds of infection with soil-transmitted helminths (odds ratio (OR) 0.66, 95% confidence interval (CI): 0.57–0.76) (Strunz et al. 2014). Sanitation access is considered a primary barrier to infection, by excluding disease-causing microorganisms from the environment (Freeman et al. 2017). However, poor maintenance of sanitation facilities can turn these into re-infection points (Campbell et al. 2015) which could be a major issue in a shared facility such as community ablution blocks (CABs).

Furthermore, access to water within informal settlements or slums is a major challenge. For instance, approximately 32% of inhabitants within informal settlements in South Africa had access to piped water in 2015 (Mutyambizi et al. 2020). The rest of the households either relied on community-shared water supply or had no access at all. The lack of access to water could potentially increase the risks of infections due to its impact on hygiene. Hygiene has been implicated as one other major factor contributing to diseases (Lange et al. 2019). For instance, about 7.75% of total deaths in sub-Saharan Africa is attributed to unsafe water, sanitation, and hygiene practices (Zerbo et al. 2021). Handwashing has been associated with a lower transmission of ascariasis (Fung & Cairncross 2009; Bartram & Cairncross 2010). Fewtrell et al. (2005) has shown an overall protective effect of hygienic practices, with handwashing reducing diarrhoeal diseases by 54%. This practice is particularly protective against severe diarrhoea (48%) and Shigellosis (59%). Proponents of the hygiene theory have stated that handwashing is a low-cost approach for reducing diarrhoeal diseases as compared to faeces disposal (Borghi et al. 2002). Additionally, this intervention reduces acute respiratory infections (Fewtrell et al. 2005).

This paper, therefore, presents a study on the role of sanitation and hygiene in diarrhoeal infection among inhabitants of informal settlements in the city of Durban, eThekwini Municipality of South Africa. A combination of secondary (retrospective) and primary data was used to determine the association between sanitation access, mainly the use of CABs, hygiene practices, water access, and solid waste management, and diarrhoea.

Two sources of data assigned as primary and secondary were used for this study. The secondary data contain diarrhoeal cases reported in the primary health care (PHC) clinic servicing the study area. The primary data were collected to provide first-hand diarrhoeal data, since not all diarrhoeal cases could be reported at the clinics.

Recorded cases of diarrhoea at a PHC clinic serving the two informal settlements were used. Diarrhoeal cases recorded between 2010 and 2017 were considered for this study.

Drone mapping was conducted at both sites in August 2019 using a quadcopter drone (dji mavic pro; DJI, USA), driven by a trained professional drone pilot. This was to generate a real-time mapping of the settlements to allow for a better sampling approach, via the division of each settlement into quadrants for representativeness. Within the settlements, 66 and 44 households were interviewed in each quadrant, respectively, for the Kennedy Road and Foreman Road informal settlements. Therefore, a total of 264 households were interviewed in the Kennedy Road informal settlement and 176 in the Foreman Road informal settlement. One person was interviewed in each household, six times during the study, resulting in total responses of 440. The diarrhoeal data were collected six times during the study, with a 100% response rate. The first household was chosen based on willingness to participate; thereafter every third household was selected until the required number per quadrant was reached. If participants in a chosen house were not willing to participate, the nearest dwelling was chosen. The location of each household added to the study was captured with a Trimble Juno 3D portable handheld Global Positioning System (GPS) data collector and the data was downloaded with the GPS Path Finder Software (https://geospatial.trimble.com/products-and-solutions/gps-pathfinder-office). Within each case study area, 264 (Kennedy Road) and 176 (Foreman Road) completed responses were achieved. The diarrhoea collection tool was administered verbally by trained bilingual fieldworkers for 6 months from September to November 2019 and January to March 2020 using the tool attached in the Supplementary Material, Appendix I. A 7-day recall period was used to record the incidence of diarrhoea during the study period, with diarrhoea defined as the passage of three or more loose stools per day (WHO 2017). Due to a nationwide lockdown in March 2020, the final week of diarrhoeal case recording (24–30 March 2020) could not be undertaken. Additionally, household questionnaires were administered in August 2019 to the selected households to ascertain their use of the CABs, hygiene practices, water access, and solid waste management among others (Supplementary Material, Appendix II). Using the GPS data collected for the location of the households and the CABs, the spatial relationship between diarrhoeal cases and the location of CABs was determined. This analysis was done by taking into consideration the data on the households with high diarrhoeal cases and their distance to the CABs in the community.

Descriptive statistics were performed with SPSS 26.0 software (IBM, USA). The difference in total diarrhoeal cases reported in the two settlements was compared using a two-tailed Mann–Whitney test with a CI of 95%. Total diarrhoea was calculated and the difference in the number of cases for the different age reporting categories of diarrhoea was determined using the Kruskal–Wallis test with a 95% CI. The prevalence odds ratio (POR) to determine the association between the data captured with the questionnaires and the likelihood of diarrhoeal infection was determined using logistic regression (Altman 1991).

This project received ethical approval from the University of KwaZulu-Natal (UKZN) Biomedical Research Ethics Committee (BE339/19). Before conducting a survey with a respondent, a written informed consent was obtained.

As captured in Table S1 (Supplementary Material, Appendix III), females dominated as respondents in this study (234 respondents). The average number of diarrhoeal cases was higher in the female population (15.7 (±15.3) cases) as compared to the male population (10.4 (±9.6)% cases). These differences were not statistically significant. Gender-based prevalence of diarrhoea did not show a high magnitude of difference although the prevalence was high in females (3.9 (±3.6)%) compared to males (3.5 (±2.8)%).

A majority of the respondents with diarrhoea (50.8%) resorted to self-medication when infected. Self-medication was related to the educational background of the respondents. For instance, respondents with grade 12 education (69.6%), certificate or diploma (67.5%), undergraduate degree (68.4%), and postgraduate degrees (66.7%) were the most likely to resort to self-medication. These differences in the use of self-medication were statistically significant (p ≤ 0.05).

The prevalence of diarrhoea varied among inhabitants of the two settlements due to their use of the CABs, household hygiene behaviour, household water access, solid waste disposal, and physical conditions of their households. Use of the CABs and open defaecation were observed to be associated with an increased diarrhoeal infection, resulting in PORs of 1.7 (95% CI: 1.05–2.26) and 1.8 (95% CI: 0.9–3.12), respectively (Table 1). However, the majority of the inhabitants (94.4%) in the study areas relied on the CABs as their primary sanitation facility as captured in Table 1. Lack of proper hygiene was not seen to be associated with increased diarrhoeal infections based on the PORs calculated, because the highest POR was observed among respondents who stated, ‘All the members of my family wash their hands after defaecation’.

The use of communal standpipes as a source of water was associated with a high diarrhoeal infection, with POR of 1.21 (95% CI: 1.04–3.12) and the use of the collected water without treatment significantly associated with diarrhoea (1.5; 95% CI: 1.02–3.45). Additionally, respondents who relied on the municipality (or their engaged contractors) to collect their solid waste were more likely to be infected with diarrhoea, with a calculated POR of 1.8 (95% CI: 0.6–2.69). Households with faecal material in proximity were significantly more likely to be infected with diarrhoea, with PORs of 1.69 (95% CI: 1.25–3.10).

In the current study, the retrospective data showed that the monthly prevalence of diarrhoea ranged from 0.75 to 2.3%. However, the primary data collected over the 6 months reported a monthly prevalence between 2.49 and 18.07%. According to a previous estimate, diarrhoea accounts for 19% of deaths in children under 5 years of age in South Africa and 9.2% globally (Kapwata et al. 2018; Ikeda et al. 2019). It is the third leading cause of death, after HIV, and low birth weight (Kapwata et al. 2018; Ikeda et al. 2019).

The high number of diarrhoeal cases observed in this study for users of the CABs could be due to several reasons. Generally, within the Durban settlements themselves, there are no sewers because private or household level toilets are non-existent. These Durban communities relied on the CABs as the only means of sanitation. Therefore, the large number of people relying on this form of sanitation (94.4% of respondents) could potentially mean high user frequency. This could potentially lead to continuous contamination of the surfaces because these CABs are only cleaned once a day, in the morning. Despite the cleaning, these surfaces could still be contaminated if the cleaning is not done thoroughly with detergents and disinfectants. It is worth noting that the efficiency of cleaning was not assessed in this study. Contact with infected individuals increases the spread of diarrhoea which is especially common in crowded areas such as informal settlements. This person-to-person (interpersonal) mode of transmission was reported by Ferrer et al. (2008) to account for 29% of all diarrhoeal cases in the city of Salvador in Brazil. The use of shared sanitation (CABs) in these informal settlements further contribute to the spread of infections and possibly the high incidence of diarrhoea. Contamination of contact surfaces in these CABs by infected individuals may be the main route of exposure. A wide variety of human pathogens, such as Enterobacter amnigenus, Enterobacter cloacae, Enterococcus spp., Shigella spp., Klebsiella spp., Escherichia coli, and Staphylococcus aureus have been found on contact surfaces in public restrooms (Burton et al. 2011; Flores et al. 2011). These pathogens can survive for varying amounts of time depending on the characteristics of the individual microorganism and environment (Abad et al. 1994; Kramer et al. 2006; Mkrtchyan et al. 2013).

Despite the observed relationship between the use of the CABs and diarrhoea in the current study, no association between the distance of the CABs and households with high diarrhoea was recorded. Other factors could have contributed to the prevalence of diarrhoea within the current study area, such as access to water, hygiene practices, or solid waste management. In the present study, the use of a communal standpipe was identified as one of the predisposing factors for diarrhoea (Table 1). This water, from the communal standpipe, is post chlorinated, treated water, supplied by the municipality and therefore can be assumed to be safe for consumption. Treated water within the informal settlements is usually collected from a communal standpipe and stored for various domestic purposes, such as drinking, cooking, cleaning, and bathing. Therefore, the water storage or collection containers could be the source of the infection. This was supported by findings in a parallel study conducted in the same informal settlements in Durban by exchange students from the Norwegian University of Science and Technology (Alesund, Norway), who found that water from these communal standpipes was not contaminated, compared to 45% of water stored within the homes (Sjåvåg & Øyhus 2019: unpublished data). The study by Sjåvåg & Øyhus (2019) identified, through microbial source tracking (MST) found that about 31% of the contamination was due to human faecal contamination. In addition to human faecal pollution, contamination by animal faecal matter could be an additional source of contamination. The findings that contamination of stored water within the homes could be from both human and animal sources highlights the possible role of hygiene practices within the home setting as a contributory factor to diarrhoeal infections. For instance, it was observed in the current study that some of the households did not have separate containers for storing water intended for hand washing or cleaning. In some instances, the water storage containers were not covered, which could easily contaminate the water.

The current study established that exposure to solid waste and faecal matter around the households were all associated with increased diarrhoeal infections (Table 1). Therefore, the findings show that access to sanitation alone may not provide a protective barrier against diarrhoea. Additional conditions such as solid waste disposal and exposure to faecal matter within the home settings could potentially be associated with increased diarrhoea.

Diarrhoea is related to socio-economic status (Gupta 2012) and therefore could account for the high number of cases reported in the informal settlements. It is estimated that South African children living in poverty are 10 times more likely to be infected with diarrhoea, than children not living in poverty (StatsSA 2012). From 2015 onwards only acute diarrhoea in children under 5 years with dehydration was recorded and is probably the reason for the lower number of diarrhoeal cases recorded from this period onwards. The management decision not to continue with the categorisation of diarrhoea into acute under 5 years, under 5 years, 5–14 years, and over 15 years may have a detrimental effect on the prevention of diarrhoea and other related diseases. The records on diarrhoeal cases, obtained in the current study, within the different age groups provide information that is critical in the determination of the most vulnerable age groups, which is crucial in the implementation of effective interventions. Some of these interventions include vaccinations, micronutrient supplements, and exclusive breastfeeding promotions (Chola et al. 2015). The age group with the highest number of diarrhoeal cases, in the current study, was children under 5 years followed closely by children aged 15 years and above. For instance, it is estimated that about 2–3 incidents of diarrhoeal cases per year are recorded for every under-5 child in 21 countries in sub-Saharan Africa and eight in Asia (Boschi-Pinto et al. 2009). The 2010 General Housing Survey (GHS) of South Africa found that over 60,000 cases of diarrhoea in children under 5 years are reported per year resulting in over 9,000 deaths (StatsSA 2012). Furthermore, Bauleth et al. (2020) found that children living in informal settlements were about 36.42 times more likely to be infected with diarrhoea compared to children who do not live in informal settlements. These two reports support the findings on diarrhoeal cases among children under 5 years in the current study.

Furthermore, this study established that the historical data obtained via data from the PHC clinics serving these communities, had different prevalence compared to the primary data collected during the period of this study. This difference could be attributed to self-medication. We observed in the current study that over 50% of the respondents resort to self-medication. Furthermore, the current study found that self-medication was common among educated respondents. The frequency of self-medication in the current study is higher than a survey in China, where about 31% of people were reported to self-medicate (Hou et al. 2013). Furthermore, Kumari et al. (2018) reported higher self-medication (81.48% of males and 85.07% of females) among medical school students in India. Findings from previous studies and the current study show that education does not reduce the practice of self-medication. Additionally, these results, indicate that clinical diarrhoeal data recorded at clinics or hospitals may not be reliable in assessing the prevalence of this disease in settings such as informal settlements.

Diarrhoeal infection, among inhabitants of informal settlements in Durban, was found to be high based on clinical records (secondary data) and primary data collected in the current study. Clinical diarrhoeal data were categorised per age groups from 2010 to 2015, with clear difference in diarrhoeal incidence among the different age groups. The variation in the number of diarrhoeal cases for the different age groups was consistent with reports from published studies, with children under 5 years being the most vulnerable. However, from 2015 to 2017 categorisation based on the age groups mentioned above was discontinued. This change in clinical diarrhoeal record keeping made it difficult to ascertain whether the same trend persisted in the last 2 years. The prevalence of diarrhoea also differed significantly between the primary data collected and the retrospective (secondary) data from records at the primary health clinic. This variation could be due to self-medication. Furthermore, the current study observed a high incidence of diarrhoea among users of the CABs in the informal settlements. However, despite the higher association between diarrhoeal infection and use of CABs, as compared to the non-users, it may be false to assume use of the CABs is the sole factor responsible for the high incidence of diarrhoea in the study areas. This is because over 90% of the respondents use these CAB sanitation facilities. The current study also found higher association between diarrhoea and hygiene practices in the households of respondents, the non-treatment of water before use, and the presence of solid waste and faecal materials around the households.

Therefore, it can be concluded that diarrhoeal incidence within the informal settlements in Durban is associated with several factors, which calls for a multifactorial approach. The provision of sanitation through the CABs should be accompanied by education on hygiene practices, water storage, and waste management within the home settings. The findings presented in this study, therefore, shows that access to improved sanitation alone may not be enough to reduce diarrhoea and related diseases.

Despite the significant findings made in this study, the following two major limitations were identified and worth noting:

• Reliance on secondary data: The study used secondary data recorded as part of diarrhoea data records by health authorities in the study area. It is worth noting that this could be a major limitation since self-medication is a common practice in the study as observed via responses in the questionnaires.

• Challenges with self-reporting of diarrhoea: The primary diarrhoea data were based on self-reported diarrhoeal cases. The limitation with such as an approach may also include research apathy or respondents not responding truthfully.

• Overestimation of prevalence ratio: The use of POR has been reported to result in the overestimation of prevalence ratio. Therefore, the association between the various factors assessed in this study could be higher than calculated due to the use of POR in this study.

The authors are deeply grateful to the respondents and their families in the settlements visited for their invaluable time with weekly interviews. We are also indebted to the community field workers for their support with data collection and logistics. The authors express their sincere gratitude and appreciation to Head and Deputy Head of eThekwini Health Department: Mrs R. Van Heerden and Mrs Zinhle Buthelezi; Mr Bev Harrod (eThekwini Health Department—Clinical Statistics) with timeous and detailed provision of clinical data. Special thanks to Mr V. Govender (UKZN School of Built Environment) and Nick Alcott, Sarah Jarman, and team (Khanyisa NGO, SA) with GIS mapping and creating of maps.

This study was funded by the Water Research Commission of South Africa (WRC Project Number: K5/2896//3).

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

The authors declare there is no conflict.