Assessing factors affecting the use of chlorine dispensers for household water treatment in rural Kenya, Malawi, and Uganda

Access to clean and safe water is important for human health and a key indicator of global development. The United Nations' Sustainable Development Goal (SDG) 6 aims to ensure access to clean water and sanitation for all by 2030 (United Nations 2021). As of 2020, 26% of the global population (2 billion people) did not have access to safely managed drinking water (UN-Water 2021), 30% of whom are in Sub-Saharan Africa. A safely managed service is defined as an improved drinking water source that is accessible on the premises, available when needed, and free of fecal and priority chemical contamination (UN-Water 2021). As of 2022, 39% of Sub-Saharan Africa population did not have access to basic water supply (World Bank 2022). Specifically, 20% of households in Kenya in 2022 (KNBS & ICF 2023), 13% in Uganda in 2016 (Uganda Bureau of Statistics – UBOS and ICF 2018), and 13% in Malawi in 2016 (NSO & ICF 2017) had no access to improved sources of drinking water, a majority of which are in rural areas. Even those with access to improved water sources are not guaranteed safe water, as an improved water source does not necessarily imply safe drinking water (Rajapakse et al. 2023). Most of the rural water sources are unprotected, leaving them exposed to animal and human contamination (Bain et al. 2014).

Drinking contaminated water can lead to water-borne health issues such as diarrhea (Solomon et al. 2020), cholera (Taylor et al. 2015), and so on. Despite the global efforts to eliminate diarrheal diseases, they remain the ninth leading cause of death among all ages and the fourth among children under five years of age (Troeger et al. 2017). In 2015, diarrhea was responsible for 499,000 deaths globally, half of which were from Sub-Saharan Africa. Particularly, diarrheal deaths of children under five years of age in 2015 were 8,900 in Kenya, 7,800 in Uganda, and 5,400 in Malawi (Troeger et al. 2017). According to the National Demographic Survey, the prevalence of diarrhea among children under five years of age was 14% in Kenya in 2022 (KNBS & ICF 2023), 20% in Uganda in 2016 (Uganda Bureau of Statistics – UBOS and ICF 2018), and 22% in Malawi in 2016 (NSO & ICF 2017).

A significant proportion of diarrheal diseases can be prevented by ensuring availability of safe drinking water and improved sanitation and hygiene. Several stakeholders are making efforts to help communities access safe drinking water. These interventions include policy making by both national governments and international organizations (Ryu 2019), development of innovative point-of-use water treatment equipment (Ehdaie et al. 2017; Okogwu et al. 2021), promotion of water treatment and safe water storage and handling practices (Ngai & Fenner 2014), and human rights advocacy (Schroering 2021). A typical community water treatment procedure includes coagulation, flocculation, sedimentation, filtration, and disinfection in successive order (CDC 2024a, 2024b). Coagulation and flocculation procedures remove turbidity and organic particles, while disinfection kills microbial pathogens in water (Rezaee et al. 2017). Chlorination is a simple, low-cost, and highly effective disinfectant in comparison to other interventions (Bänziger et al. 2022). Research has shown that point-of-use water treatment is effective in deactivating microbial contamination in drinking water, hence reducing the burden of water-borne illness (Wolf et al. 2014). A meta-analysis study indicated that point-of-use treatment of drinking water with chlorine reduces diarrhea in children by 29% (Arnold & Colford 2007). Another study review showed that point-of-use disinfection products may reduce diarrhea by 25% (Clasen et al. 2015).

The effectiveness of chlorine depends on the amount of organic and inorganic matter in water, pH and alkalinity of water, presence of microorganisms, as well as water storage and handling practices followed in the process of indoor storage (CDC 2024a, 2024b). Since chlorine does not effectively penetrate organic particles in water, coagulation is recommended before application of chlorine on highly turbid water (CDC 2024a, 2024b). A higher chlorine dose and longer contact duration are required to destroy higher microbial risks in water (Jin et al. 2020; WHO 2013). When chlorine is added to water, it reacts with organic material, and other compounds present in water and the remaining chlorine concentration is called total chlorine. Total chlorine residual comprises combined chlorine, which has reacted with nitrates in water, and free chlorine, which is available for disinfection (WHO 2013). The recommended range for the presence of free chlorine levels in water at the point-of-use is about 0.2–4 mg/L (WHO 2013; CDC 2024a, 2024b). Free chlorine levels ensure the water is protected from recontamination during indoor storage. Recontamination is introduced by dirty storage containers, accessing water through dipping, and not covering the storage container (Murphy et al. 2016; Gärtner et al. 2021). Therefore, the safety of chlorinated water requires that careful tests be conducted to produce a dosage that optimizes the health benefits of chlorination by defining an acceptable trade-off between microbial risks on the one hand and chemical risks and unpleasant taste on the other hand (Levy et al. 2014; Gámiz et al. 2020; Leonard et al. 2022; WHO 2013; CDC 2024a, 2024b).

Evidence Action is running the Safe Water Now program that provides rural communities in Kenya, Uganda, and Malawi access to safe drinking water through the installation of chlorine dispensers at communal water points. A chlorine dispenser is a calibrated equipment designed to dispense chlorine-based disinfectant required to treat drinking water at the point of use. The dispensers are strategically installed near communal water collection points for ease of access and to serve as physical reminders to the community to treat their drinking water. As of 31 December 2022, the program had served more than 4.5 million people across Kenya, Uganda, and Malawi, allowing them to improve the microbiological quality of their household drinking water at the point of use. The program utilizes a model that incorporates community-based promoters, who are volunteers responsible for maintaining the dispensers as well as conducting community sensitization on the benefits of treating water with chlorine. Safe Water Now program's chlorine dispensers are calibrated to dispense 3 mL of dilute sodium hypochlorite (chlorine), which is sufficient to disinfect 20 L of water.

Despite efforts through the Safe Water Now program to promote water treatment practices by provision of free chlorine, there still remain a significant number of households that are not treating their drinking water (non-adopters). Specifically, the program data show non-adoption rate of 37% between 2021 and 2022 across Kenya, Malawi, and Uganda. There is therefore a need to understand various factors that influence a community's adoption of chlorination. Studies have related adoption of water treatment practices to a wide range of factors including the quality of treated water, demographic characteristics, knowledge, and religious and cultural beliefs. In Ethiopia, studies have shown that gender and education level of household head, residency, ownership of radio or television, mode of accessing water, promotional training, and frequency of collecting drinking water are related to water treatment adoption rates (Belay et al. 2016; Eticha et al. 2022; Tamene et al. 2022). A review of studies in India, Kenya, and Egypt found that a child's age and gender, school grade level, socioeconomic index, and hygiene knowledge are associated with the child's health indicators such as school absenteeism, diarrhea, and malnutrition (Joshi & Amadi 2013). A study in Arizona revealed that residents with higher income and education levels were more likely to treat their water (Lothrop et al. 2015). A study in Pakistan showed that households with larger families are more likely to adopt chlorination (Akram 2020). A study in Nigeria showed that residence type, education level, and wealth index were associated with household water treatment (Azeez et al. 2023). These examples are not nearly as exhaustive, but they highlight the key adoption factors for water treatment technologies. However, there is limited literature highlighting factors for non-adoption of chlorination in the regions considered in this current study. The aim of the study, therefore, was to identify factors influencing the use of chlorine dispensers for point-of-use household water treatment in Kenya, Uganda, and Malawi.

The study was done in rural parts of Western Kenya, Eastern Uganda, and Southern Malawi. The sites were selected because of their high prevalence of under-5 diarrhea rates and the cost-effectiveness of implementing the program.

The study used secondary data from the Safe Water Now program's routine monitoring data. The routine monitoring data utilizes a repeated cross-sectional study design to assess the performance of the program in rural communities in Kenya, Malawi, and Uganda. This assessment was repeated at 2-month intervals from 1 January 2021 to 12 December 2022. A previous study showed that cross-sectional studies may be used to study associations of multiple exposures and outcomes (Wang & Cheng 2020). The assessment does not require the measurement of the exact same participants in all its iterations, and this provided the flexibility it needed without compromising its scientific integrity. The data were collected primarily for the monitoring and evaluation of the program and not for the research study.

By the end of 2022, the program had installed more than 28,000 chlorine dispensers across the study geographies, serving more than 800,000 households. All community members using either improved or unimproved water sources with a chlorine dispenser, who are above 18 years and are responsible for the household drinking water, were eligible to participate in the assessment. Similarly, promoters in charge of the water sources with chlorine dispensers were included to assess their knowledge levels and the use of chlorine.

A stratified sampling method with probability proportionate to size was used to select the water sources and households to be visited. Water sources with chlorine dispensers were randomly selected from each country, one promoter and four randomly selected households using the selected water source. A total of 17,244 households (Kenya: 9,540, Malawi: 3,826, and Uganda: 3,878) and 4,251 (Kenya: 2,398, Malawi: 917, Uganda: 936) water sources were used for the study.

Chlorine adoption was the outcome variable in the assessment, and is defined as the confirmed presence of total chlorine residuals (TCRs) in a household's drinking water. TCR is the chlorine concentration in water that remains after chlorine has reacted with organic materials and metals. To test the presence of TCR in water, a field officer adds DPD (N,N diethyl-p-phenylene diamine) reagent to sampled water – which turns pink in the presence of chlorine, places the water on a graduated color wheel comparator kit, and records the TCR readings that range in value from 0 to 3.5 mg/L. Chlorine adoption was then defined from the TCR readings, where households with TCR values greater than zero were classified as adopters, and those with TCR value of zero, as non-adopters.

The explanatory variables were as follows: a household's and promoter's socio-demographic characteristics, a household's water collection, storage and handling practices, a household's sensitization and knowledge of chlorine and dispensers, a promoter's chlorine adoption, and the status of the dispenser.

Socio-demographic characteristics include family size, presence of a child under five years of age in the household, as well as age, gender, and education level of the promoter. The age and gender were used as indicators of a promoter's level of active involvement in performing their duties, while the ability to read and write was considered as a key factor for a promoter's comprehension of the water treatment procedure, for which basic primary education sufficed.

Household's water collection, storage, and handling practices include the age of the person collecting water, nature of the water storage container, way of accessing drinking water, cover type of the storage container, and when the water was collected. Observance of good storage and handling practices is a composite of water storage and handling practices variables. The variable is categorized as ‘yes’ if drinking water is stored in a clay pot for not more than 48 h or a plastic container for not more than 72 h, the storage container is covered, and if access to drinking water from the storage container is by pouring or via a tap.

Household sensitization and knowledge variables consist of promotional activities, knowledge of the benefits of chlorine, and knowledge on the use of dispensers. Promotional activities variable was obtained by asking respondents whether they interacted with the community member in the 30 days preceding the survey. Knowledge on the dispenser use includes correct procedure of operating the dispenser and observance of a 30-min wait time before using chlorinated water. The correct procedure of operating the dispenser is cleaning the water container, adding liquid chlorine, and waiting for 30 min to drink the water. The score of the knowledge of the benefits of chlorine was derived from a question asking participants to list the benefits of chlorine. A score was then generated by dividing the number of listed benefits by the expected number of responses.

Promoter chlorine adoption is when a promoter uses chlorinated drinking water as confirmed through TCR tests at their own households.

Functional chlorine dispenser variable is assessed by turning the valve to see whether it releases liquid chlorine solution. Dispenser dysfunctionality is usually a maintenance problem resulting from hardware problems or chlorine running out.

The data were analyzed using both descriptive and logistic regression methods. Simple logistic regression was used to assess the individual effects of each variable, and those with p-value <0.25 were identified and included in the multivariate logistic regression analysis (Bursac et al. 2008). Regression diagnostics were conducted to ensure the estimates were reliable and precise. Multi-collinearity of variables was identified using variance inflation factor (VIF), and variables with VIF of more than 10 were either excluded or combined with other variables. Goodness-of-fit of the model was also checked using the log likelihood chi-square test and the proportion of correctly classified values. Statistical significance was considered at p-value < 0.05. All analysis was conducted using Stata version 18.0.

The data used in this assessment was collected as secondary information and part of the routine monitoring for the program. Clearance was obtained from the county governments and local leaders. All the respondents to the monitoring data were above 18 years and therefore consent was sought for each and every participant. It is worth noting that no personal information was collected. Data retrieved for analysis were handled with a lot of care and restricted to a limited number of people for the sake of integrity and discretion.

The assessment found that the majority (92%) of the community members interviewed were female, with a mean age of 41 years (standard deviation (SD) 13) across Kenya, Malawi, and Uganda. The average family size was six members in Uganda, and five members in both Kenya and Malawi. About half of the households in Kenya (50%) and Malawi (51%) and 63% in Uganda had at least one child under the age of five. Less than half (45%) of the promoters interviewed in Uganda were female, with 76% in Kenya and 81% in Malawi being female. The average age of the promoter was 42 years in Malawi, 49 years in Kenya, and 50 years in Uganda. The assessment further revealed that 92% of promoters in Malawi, 50% in Uganda, and 41% in Kenya had sensitized community members in the last 30 days prior to the interview. Additionally, 96% of dispensers were functional during unannounced assessment visits in Malawi, 91%, in Kenya, and 87%, in Uganda. Less than half of the community members were practicing good water storage and handling practices in Kenya (41%) and Uganda (32%), with 56%, in Malawi. In terms of water collection practices, 13% of the community members in Uganda, 9% in Kenya, and 4% in Malawi reported that children who were under 18 years of age fetched their drinking water. The chlorine benefit knowledge score was 45% in Uganda, 39% in Malawi, and 38% in Kenya. Majority of the community members in Malawi (91%) and Uganda (82%) were knowledgeable about the procedure of using the chlorine dispenser, with 57% in Kenya being aware of the procedure. Finally, 96% of the promoters in Malawi, 82% in Uganda, and 81% in Kenya were treating their water using chlorine from chlorine dispensers. The community chlorine adoption was 90% in Malawi, 60% in Kenya, and 58% in Uganda.

Multiple logistic regression results are presented in Tables 1–3, respectively, for Kenya, Malawi, and Uganda. These results show that households practicing good water storage and handling practices were between 1.25–1.83 times more likely to display positive results for chlorine residuals in their drinking water than the corresponding group (Kenya: adjusted odds ratio (AOR): 1.25[1.13–1.38], Malawi: AOR: 1.83[1.36–1.46], Uganda: AOR: 1.34[1.15–1.57]). Households where children under 18 years of age collected drinking water were 0.65, 0.42, and 0.17, less likely to display positive results for chlorine residuals in Kenya, Malawi, and Uganda, respectively, than when a person of 18 years and above collected water (Kenya: AOR: 0.65[0.55–0.77], Malawi: AOR: 0.17[0.11–0.26], Uganda: AOR: 0.42[0.34–0.58]). Households knowledgeable about the benefits of using chlorine to treat drinking water were more likely to display positive results for chlorine compared to the corresponding group. Particularly, for every one-score change in diarrhea benefits score, the log odds of using chlorinated water increased by a factor of 2.38 in Kenya (AOR: 2.38[1.81–3.12]), 3.89 in Malawi (AOR: 3.89[1.54–9.84]), and 2.14 in Uganda (AOR: 2.14[1.50–3.05]). Similarly, households that knew the correct process of chlorinating water (both use of the chlorine dispenser and to wait at least 30 min before drinking) were 1.49 times in Kenya and 2.88 times in Malawi more likely to display positive results for chlorine residuals than those who were not aware (Kenya: AOR: 1.49[1.35–1.64], Malawi: AOR: 2.88[1.96–4.23]). There was however no statistical significance of knowledge on dispenser use and chlorine adoption in Uganda. Households that received promotional information in the preceding 30 days were between 2.09–5.44 times more likely to have chlorine residuals in their water than the corresponding group (Kenya: AOR: 2.09[1.88–2.31], Malawi: AOR: 5.44[3.74–7.91], and Uganda: AOR: 2.57[2.22–2.96]). Family size was also a significant factor in Uganda. Specifically, for every one-member change in the household size, the log odds of using chlorinated water increased by a factor of 1.03 (AOR: 1.03[1.00–1.06]).

The assessment further revealed that having functional dispensers – dispensing chlorine when turned – increases the likelihood of chlorine adoption by community members (p < 0.01); particularly, households using functional dispensers were between 2.32–9.27 times more likely to display positive results for chlorine residuals in their drinking water than the corresponding group (Kenya: AOR: 2.32[1.94–2.76], Malawi: AOR: 9.27[4.05–21.21], and Uganda: AOR: 3.31[2.54–4.31]).

Regarding a promoters' capacity to influence their communities, results show that a promoter's education level was significant in Kenya. Specifically, households whose water points managed by promoters with at least a primary level education were 1.26 times more likely to display positive results for chlorine residuals than those whose promoters had no education (AOR: 1.26[1.02–1.58]). The assessment established that households whose water points were managed by promoters who themselves use chlorine from the chlorine dispensers were between 2.04–4.66 times more likely to display positive results for chlorine residuals than those whose promoters did not use chlorine (Kenya: AOR: 3.12[2.76–3.54], Malawi: AOR: 4.66[2.47–8.79], and Uganda: AOR: 2.04[1.63–2.56]).

This assessment revealed that households that observe good water storage and handling practices tend to have their drinking water display positive results for chlorine residuals. This is in agreement with a study in Uganda that recommended storing water in properly cleaned containers to reduce the amount of organic matter that reacts with chlorine hence reducing its effectiveness (Gärtner et al. 2021). Poor water handling practices may negate benefits that would otherwise be derived from treated water (Levy et al. 2014; Figueroa & Kincaid 2020), as this introduces bacteria in the treated water leading to recontamination. A study in Kenya showed that chlorination is more effective in plastic than in clay containers because of lower chlorine demand in plastics, permitting chlorinated water to remain significantly less contaminated for up to 72 h (Murphy et al. 2016). Therefore, it is recommended that households should store water in a clay pot for no longer than 48 h or in a plastic container for no longer than 72 h; the container should always be covered and drinking water accessed via a tap or by pouring.

The assessment also revealed that a child collecting drinking water is associated with a negative impact on their household's chlorine adoption. A possible explanation is that children might not have adequate knowledge to operate the chlorine dispenser. Additionally, community knowledge on the benefits of treating water was found to increase the likelihood of chlorine adoption across the three countries. It appears that the more communities become aware of water-borne diseases – their causes and prevention, they naturally become inclined to make the decision to adopt water treatment practices. A similar observation was made by studies that identified lack of knowledge about the benefits of water treatment as a barrier to sustained water treatment (Kraemer & Mosler 2010; Figueroa & Kincaid 2020). Additionally, community knowledge on how to operate dispensers was significantly associated with household water treatment. It was also observed that operational knowledge of the chlorine dispenser, correct dosage, and correct latency required before drinking chlorinated water are significantly associated with household water treatment. These observations suggest that knowledge is a key determining factor in chlorine adoption – those who understand the importance of drinking safe water and have adequate practical knowledge of accessing and handling it will make the effort to disinfect their drinking water. These findings underpin the importance of sensitization campaigns in mobilizing communities to adopt water treatment practices. The present assessment showed that community promotional activities in whatever frequency within a month generally impact chlorine adoption positively. The assessment further established that community promoters become more effective in influencing their communities when they have at least a basic level education, and when they themselves have adopted chlorination. This observation is consistent with that of a literature review that concluded that personal advocacy reinforces a person's health behavior because people are more likely to learn through observing the behavior of others, and through interpersonal and group communication (Figueroa & Kincaid 2020). Another study showed that providing training to households on how to treat their water is essential to enhancing a community's uptake of water treatment practices (Eticha et al. 2022).

Households with larger family sizes were more likely to treat their drinking water in Uganda. This was not significant in Kenya and Malawi. This result is consistent with a study in Ethiopia, where one additional household member led to higher odds of better drinking water sources (Desye et al. 2022). This may be a result of family members sharing out responsibilities, thereby relieving the household's primary caregiver of the burden of collecting and treating their drinking water. By contrast, a negative relationship was found between increased household size and water treatment in previous studies (Gaffan et al. 2022).

A functional chlorine dispenser is a significant determinant of chlorine adoption among community members. Dysfunctional dispensers could dissuade the community members, who may opt for alternative treatment options.

Owing to a relatively low uptake of water treatment practices among the rural population, this assessment measured several factors that influence their adoption of chlorine dispensers for the treatment of household drinking water at the point of use, with a specific objective of deriving information that will aid in developing effective strategies to guarantee increased and sustained uptake of chlorination.

The findings show that observance of good water storage and handling practices, adults collecting the drinking water, knowledge of the benefits of chlorinated water, knowledge of dispenser use, continued community promotional activities, larger family size, functional dispenser, educated promoters, and promoters using chlorinated water were positively associated with household chlorine adoption.

In light of these findings, the following recommendations are forwarded: (1) ensure the chlorine dispensers are functional and adequately supplied with chlorine. Specifically, ensure dispenser valves are functional and dispensing precise amounts of chlorine, (2) encourage promoters to continue sensitizing and educating the community at least once a month on the importance of chlorination and its correct procedure, (3) educate households on good water storage and handling practices, (4) extend to children under 18 years of age the water treatment education programs such as proper chlorine dosage and the procedure for using the chlorine dispenser, and (5) elect promoters with at least a primary level education and who are themselves adopters of chlorination.

The authors would like to thank all the program and field workers across Kenya, Malawi, and Uganda for the implementation of the Safe Water Now program. Appreciation also goes to the monitoring, learning, and evaluation field operations and data management team for the data collection and processing. Special thanks goes to the National, County, District Governments and local communities in Kenya, Malawi, and Uganda for the immense support accorded to the program delivery. To the promoters who played a key role in sensitization of communities on chlorine uptake, the authors remain forever indebted.

Safe Water Now program implementation and monitoring activities were funded by Astellas Global Health Foundation, Open Philanthropy Project Fund, Ray and Tye Noorda Foundation, and Effektiv Spenden Schweiz.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.