https://orcid.org/0000-0001-6412-1483
Figures
Abstract
Major investments in water supply infrastructure will be required to achieve Sustainable Development Goal (SDG) 6. Safely managed water services are also central to global cholera and diarrhoeal diseases prevention strategies. However, evidence remains scarce on how to efficiently improve piped water services in complex settings where infrastructure investments are most needed. We conducted a process evaluation of a large-scale water supply infrastructure improvement programme in Uvira, Democratic Republic of the Congo, in parallel to a pragmatic trial. Considering three evaluation domains–context, implementation, and population response–, we assessed the validity of the programme’s theory of change and underlying assumptions. Information sources included construction works documentation, operational and billing records from the water utility, and household surveys. The evaluation covers the period 2014–2021. Trial results are not within the scope of this manuscript. The programme did not achieve expected improvements in the water supply service during the evaluation period. Out of 16 assumptions underlying the theory of change, six remained valid, seven partially valid, and three turned out to be invalid. Contextual challenges included extreme flooding in 2020 and the Covid-19 pandemic, which disrupted construction works. Issues related to electricity supply and the rise of Lake Tanganyika emphasise the need for cross-sectoral approaches and consideration of climate change in the planning of water supply infrastructure. Implementation challenges underlined the importance of capacity strengthening alongside infrastructure improvements. Population response elements suggest that affordability and informal practices such as tap sharing should be taken into account. The programme was a good example of early engagement with researchers and provides unique insights into the implementation of large-scale infrastructure improvements in a complex, low-income setting. Pragmatic evaluation approaches should be adopted for the generation of scientific evidence from complex programmes in order to optimise future infrastructure investments contributing to progress towards SDG6.
Citation: Gallandat K, Hutchins C, Malembaka EB, Jeandron A, Saidi JM, Rumedeka BB, et al. (2024) Process evaluation of an urban piped water supply infrastructure improvement programme in Uvira, Democratic Republic of the Congo. PLOS Water 3(10): e0000185. https://doi.org/10.1371/journal.pwat.0000185
Editor: Francis Dakyaga, SD Dombo University of Business and Integrated Developmental Studies, GHANA
Received: August 13, 2023; Accepted: September 17, 2024; Published: October 31, 2024
Copyright: © 2024 Gallandat et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data used to generate Fig 4 are available on the following public Open Science Framework (OSF) folder: https://osf.io/76sev/?view_only=e96559353c004ca7ad048b68c42cd380 All household survey data used for preparation of the manuscript will be made available upon publication on the following OSF folder: https://osf.io/tdspg/?view_only=f85189ea00fd42b1b58ab81ec409c8ae All other documents and datasets referenced in the manuscript are available upon request from the corresponding author.
Funding: This study was funded by the French Agency for Development (AFD, ref. EVA/364-2015) and the Veolia Foundation (VF, ref. 13/14 HD 1123), with KG and OC as Principal Investigators. AFD and VF reviewed and provided inputs for this manuscript, with AS (AFD), PYD (AFD) and DM (VF) as co-authors. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: I have read the journal’s policy and confirm that the authors of this manuscript have the following competing interests to declare: Audrey Seon and Pierre-Yves Durand are employed by AFD and Damien Machuel is employed by VF; AFD and VF co-financed the water supply infrastructure improvement programme in Uvira. The other authors have no conflicts of interest to disclose.
Introduction
Large-scale investments in water supply and sanitation infrastructure will be necessary to achieve universal access to safely managed services–Sustainable Development Goal (SDG) #6 –by 2030: an estimated US$ 114 billion per year would be needed [1]. While current financing levels are much lower, substantial investments in water supply and sanitation infrastructure are expected in low-income, urban areas in the next decade. Optimising these investments is critical: an estimated 38% of current global infrastructure investments (including water and other sectors) are not spent effectively [2].
Improvements in water, sanitation and hygiene (WASH) are fundamental to prevent diarrhoeal diseases, including cholera, which remain a leading cause of morbidity and mortality. Globally, an estimated 0.93–1.16 million deaths due to diarrhoeal diseases were attributable to inadequate WASH access in 2019 [3]. Cholera alone causes 1.3–4.0 million cases and 95,000–143,000 deaths per year, according to modelling estimates that account for the widespread under-reporting of cases [4]. The WHO has declared a global surge in cholera with 44 countries reporting cases in 2022, including very large outbreaks (>100,000 cases) in seven countries and across two regions (Africa and Asia) [5]. An epidemiological analysis of cholera cases recorded across sub-Saharan Africa from 2010–2016 showed that over half of the cholera burden was geographically concentrated in endemic districts–or hotspots–that represent less than 4% of the population [6]. Populations with limited WASH access are disproportionately affected [7, 8]; a recent study focusing on sub-Saharan Africa suggests that a 1% increase in access to a piped or improved water supply is associated with a 3.5% and 7% decrease in cholera incidence at district and country level, respectively [9].
The roadmap of the Global Task Force on Cholera Control (GTFCC) to ending cholera by 2030 calls for endemic hotspots to be prioritised for the implementation of long-term WASH improvements [10], in line with SDG #6. However, experience with, and documentation of, interventions to increase access to high-quality, piped water services remains limited in the fragile, challenging settings where cholera is most prevalent. A recent systematic review of WASH interventions effectiveness suggested that the supply of higher quality water on premises could reduce diarrhoeal disease risk by 52% (95% CI: 13–74%) compared to an unimproved water source but this estimate was based on two studies only, including one assessing an education intervention in rural communities of Puerto Rico and another focused on in-line chlorination [11]. None of the studies identified in the systematic review was conducted in an emergency setting and none focused on acute diarrhoeal diseases or cholera.
Monitoring a high quality of water supply service in line with SDG6 definitions requires assessing multiple dimensions including accessibility, water availability and quality, as per guidance provided by the Joint Monitoring Programme (JMP) [12]. Water supply infrastructure improvements can be complex interventions, whose effects depend on a range of context-specific elements (e.g. climate, natural water resources, economy, social and cultural norms), especially when implemented in low-resource, volatile environments like many cholera hotspots [13, 14]. It is therefore important for quantitative impact evaluations of such interventions to be complemented with different approaches such as process evaluations in order to capture learnings, to better understand mechanisms at play and to identify factors influencing their outcomes [15, 16].
This manuscript presents a process evaluation of large-scale water supply infrastructure improvements in the town of Uvira, an endemic cholera hotspot in Eastern Democratic Republic of the Congo. This process evaluation was conducted in parallel to a pragmatic health impact evaluation of the infrastructure improvement programme, including a trial and an economic evaluation [17].
The context of our study, with limited access to basic services, protracted conflict, frequent population displacement, and a high vulnerability to natural disasters, is characteristic of the challenges faced in high-burden, cholera-endemic areas. The main goals of the process evaluation were to identify lessons learned applicable to future infrastructure improvement programmes in similar settings and to put health impact evaluation results into perspective. The process evaluation draws from project documents, utility records and household surveys to complement the trial findings by systematically investigating whether and how context, implementation modalities, and population response may enable or hinder piped water service improvements. This work addresses important research gaps by investigating the implementation of a large-scale, urban, piped water supply infrastructure improvement programme (a type of WASH intervention that has rarely been studied [11]) in a low-resource, complex emergency setting (a type of context that is highly relevant for future investments but where such studies have not been carried out).
Methods
Study setting
Uvira is a town of approximately 280,000 inhabitants (according to 2020 official records) in South Kivu, DRC. It is located at an altitude of approximately 800 m above sea level, on mountainous terrain on the shore of Lake Tanganyika, and has a moderate tropical climate. Five rivers cross the town of Uvira, which is prone to flooding. Protracted conflict and continuing population displacements have been affecting the town and surrounding region for almost three decades [18, 19].
Diarrhoeal diseases and cholera are endemic in Uvira, which is located within a so-called cholera “hotspot” and recognised as an endemic area [6, 20]. An annual average of 1,200 admissions (range: 451–1883) was recorded between 2009 and 2021 at acute diarrhoeal disease treatment facilities managed by the Uvira Health Zone and where data were collected for the health impact evaluation of water supply improvements. Variability was observed between year and seasons, as detailed elsewhere [17]. The use of rapid diagnostic tests since 2016 suggests that approximately 40% of patients admitted to cholera treatment facilities are cholera cases [17, 21].
The Uvira water supply network was built in 1958 and severely damaged during the First Congo War (1996–1997) [22]. Before the start of the infrastructure improvement programme evaluated here, a preliminary assessment conducted by the French Agency for Development (AFD) suggested that approximately 30% of the Uvira population had access to an intermittent piped water service [22]. Household surveys conducted in 2016/17 suggested that a majority of households (65%) had to collect water from multiple sources and that rivers (62%), taps outside the compound (53%), and Lake Tanganyika (34%) were the most commonly used [23]. Very few households (6%) reported treating water; settling or cloth filtering were the main treatment methods and free residual chlorine was detected in 8% of household water samples, primarily collected from the piped water supply system. There was no centralised wastewater management system; households reported using shared compound latrines (41%), private latrines located outside (27%), or practicing open defecation (20%).
Infrastructure improvement programme
In 2014, AFD and the Veolia Foundation (VF), in collaboration with the national publicly owned company REGIDESO S.A., initiated a project to improve the water supply infrastructure in Uvira. The water supply system consisted of a main water intake on the Mulongwe river connected to a water treatment and pumping station, a 1,600-m3 storage tank uphill in the southern part of the city, and a piped distribution system with 3,150 household connections [22]. The project included an upgrade and doubling of the capacity of the main water treatment plant (with coagulation/flocculation, sand filtration, chlorination) and pumping station, as well as the construction of a new 2,000 m3 tank located in the northern part of the city. Approximately 24 km of new pipes were installed and 10 km of existing pipes were rehabilitated. It was initially planned to build 115 new community taps and to install or rehabilitate 2997 private taps, allocated to different areas (‘clusters’) in the city based on population density, network coverage–including length of existing, rehabilitated, and new pipes–, and cholera incidence as estimated by the study team based on clinical surveillance. Between September 2019 and December 2021 (end of the period of observation for our study), 56 new community taps were built, 1191 new private connections were installed and 717 existing private connections were rehabilitated. The actual number of private taps installed in each cluster was demand-based and adapted to respond to an evolving context (e.g. the lake level rise made installation impossible in some areas and users’ demand exceeded planned quotas in others). REGIDESO was responsible for informing and promoting new private connections among the population. A mandate was given by REGIDESO to the NGO ADIR following a call for proposals to support the establishment of management committees and development of sustainable financial models for the operation of community taps. The project was extended beyond the timeframe of the study and is still ongoing; considering works completed after December 2021, a total of 2368 households taps were installed or rehabilitated, and 93 community taps were built.
Main study
We refer to a health impact evaluation of the infrastructure improvement programme commissioned by AFD, VF and REGIDESO S.A. commissioned as the “main study” [17]. A Memorandum of Understanding was signed in 2015 between AFD, VF, REGIDESO, the DRC Ministry of Health, OXFAM International in DRC, and the London School of Hygiene and Tropical Medicine (LSHTM). A stepped-wedge cluster randomised trial was designed specifically to assess the effects of water supply service improvements on cholera and diarrhoeal diseases [24]. The city was divided into 16 clusters of varying areas (0.36–2.75 km2) and population sizes (3,996–56,106 in 2018) based on the length of pipes to be rehabilitated or installed and the number of new community taps, as well as consideration of administrative and natural boundaries (Fig 1). The main outcome for the trial was monthly cholera incidence, as estimated based on clinical surveillance at cholera treatment facilities managed by the Uvira Health Zone. We estimated associations between suspected cholera incidence (based on admissions to cholera treatment facilities) and drinking water service quality (RR 0·86, 95% CI 0·73–1·01), the quantity of drinking water consumed (RR 0·80, 95% CI 0·62–1·02) and the continuity of drinking water services (RR 0·81, 95% CI 0·77–0·86). We also estimated associations between confirmed cholera incidence (i.e. including only patients with a positive rapid diagnostic test result) and drinking water service quality (RR 0·84, 95% CI 0·73–0·97), quantity (RR 0·76, 95% CI 0·61–0·94) and continuity (RR 0·75, 95% CI 0·69–0·81). These results suggest that ensuring a sufficient and continuous piped water supply may substantially reduce the burden of endemic cholera and diarrhoeal diseases [17].
Basemap from OpenStreetMap: https://www.openstreetmap.org/search?query=uvira#map=13/-3.38354/29.13394.
Cross-sectional household surveys were conducted between 10-25th October 2016; 25th October -11th November 2017; 9-12th July 2020; and, 18-30th September 2021 with a focus on water use and related practices.
Process evaluation
The process evaluation was designed to complement the main study. It was based on theory [15, 16, 25, 26] and inspired from experiences with similar work [27, 28]. In line with Medical Research Council guidance [15], we started by defining a theory of change (ToC) describing the impact mechanisms of the evaluated intervention and assumptions underlying causal links, which was published in the main study protocol and served as a guiding framework [24].
The ToC (presented in Fig 2 with all original elements, and additions specific to this process evaluation) outlined how a reduction in diarrhoeal disease and cholera incidence (the desired impact of the intervention) could theoretically be achieved through water supply infrastructure improvements (programme inputs) which would increase the quantity and improve the quality of water supplied through the piped network in Uvira (intermediary outcomes) [24]. Combined with the uptake of safe water storage and hygiene practices at household level, these improvements to the water supply service would reduce the risk of diarrhoeal disease transmission (outcome). The ToC included 16 assumptions underlying causal chain links from inputs to impact (Fig 2, Table 1).
Added elements compared to the original theory of change include (i) traffic light colours for the validity of assumptions (green: valid; orange: partially valid; red: invalid); (ii) brief justifications for the assumptions’ validity assessments in coloured text; (iii) red signs for broken causal links.
The ToC assumptions were categorised according to three domains considered in the process evaluation, based on the Context and Implementation of Complex Interventions (CICI) framework [16]: context, implementation of the intervention, and population responses (Table 1). For each assumption and domain, guiding questions were developed and potential sources of information identified (S1 Table). We then sought to identify elements that affected (i) all programme activities, (ii) the water supply service or intervention, and (iii) the impact evaluation or research activities. The validity of each assumption was assessed qualitatively based on all available, relevant information, following guiding questions. Whenever possible, objectively measurable indicators (e.g. descriptive statistics from household surveys or operational records) were used to address guiding questions. In addition, unexpected events or effects that were known or discovered by the research team were documented and allocated to one of the three domains. Results were discussed and consulted with all co-authors (representing the research, works supervision, and donor teams) until reaching consensus based on available evidence.
Data collection and analysis
Data sources included: documents shared by the programme team regarding construction works implemented as part of the programme (used to assess ToC assumptions #3, 5, 7, 15 (Fig 2, Table 1));; REGIDESO operational records including hours of pumping with electricity from the network or generators and volumes of water produced (used to assess ToC assumptions #6, 10); billing records shared by REGIDESO for domestic users connected to the water supply network, and by ADIR documentation for community taps (used to assess ToC assumptions #5, 7, 10); household survey reports (used to assess ToC assumptions #1, 5, 7, 9, 10, 11, 12, 13); data from the main study (used to assess ToC assumptions #3, 12, 14); and, programme documents including reports, minutes of meetings, and emails (used to assess ToC assumptions #1, 2, 3, 4, 8, 15). Descriptive statistics were calculated and graphs prepared using Microsoft Excel (2019) for selected household survey and REGIDESO operational data. For example, in order to assess the validity of assumption #6 about reliability of the electricity supply, the daily average hours of supply by the national company SNEL were estimated based on operational records provided by REGIDESO and, in order to assess the validity of assumptions #11 about hygiene behaviours, survey data about handwashing materials and times were analysed.
The period considered for the process evaluation was from initiation of the project in 2014 until end of December 2021 (Fig 3); construction works implemented from 2022 onwards due to delays and/or new project components are not captured in this work.
The theory of change was adapted from the version published previously [24] using software Draw.io (v18.0.7) to show where assumptions did or did not hold true and where the causal links may have been disrupted.
Ethics approvals
Household surveys mentioned in this manuscript were conducted as part of the main study. All data collection activities were approved by the ethics committees at the London School of Hygiene and Tropical Medicine (ref. 8913) and at the School of Public Health, University of Kinshasa, Democratic Republic of the Congo (ref. ESP/CE/088/2015). Written informed consent was obtained from all participants and survey questionnaires were only administered to individuals aged 18 years or older.
Inclusivity in global research
Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information (S1 Checklist).
Results
Out of 16 assumptions underlying the theory of change, six remained valid, seven partially valid, and three turned out to be invalid (Table 1, Fig 2). Valid assumptions were within the implementation domain (n = 3, primarily related to capacities to implement project and research activities), context domain (n = 2, focused on security and on the importance of water and hygiene for diarrhoeal disease prevention), and population response (n = 1, namely the preference for piped water if available). Partially valid assumptions were within the implementation domain (n = 3, about reliable supply chains, adequate management structures, and affordable prices for new connections), context (n = 2, linked to support to the programme by local stakeholders and research uptake), and population response (n = 2, related to hygiene practices and care-seeking behaviour). There was one invalid assumption in each evaluation domain, related to (i) a reliable electricity supply (contextual challenge), (ii) the expected increase in quantity and improvement in quality of the water delivered to households (implementation challenge); and, (iii) the local demand for new connections to the water supply network, which is closely linked to affordability issues (population response challenge). Relevant elements to support these validity assessments are summarised in Table 1 and presented hereafter in a narrative that covers each evaluation domain, distinguishing between those affecting the infrastructure improvement programme, research activities, or both.
Process evaluation domain #1: Context
The analysis of contextual elements highlighted two unforeseen events that affected both project and research activities, namely the Covid-19 pandemic and extreme flooding events in April 2020. In addition, electricity supply issues and the lake level rise primarily affected implementation of the intervention, whilst the deployment of a mass oral cholera vaccination (OCV) campaign in 2020 had implications for the research.
Following declaration of the Covid-19 pandemic by the World Health Organisation (WHO) on 11th March 2020, Burundese and Rwandan borders with DRC were closed on 20th and 21st March 2020, respectively. This hindered key staff from the construction team from returning to Uvira, caused substantial delays in the supply of construction materials and led to difficulties in finding a suitable workforce. With respect to the epidemiological impact evaluation, the pandemic may have affected outcomes in two ways: (i) hygiene promotion may have contributed to better handwashing practices and reduced diarrhoeal disease transmission; (ii) Covid-19-related fears may have influenced care-seeking behaviour, leading diarrhoea and cholera cases to avoid healthcare facilities.
On 17th April 2020, extreme flooding destroyed approximately 3,500 houses and affected an estimated 80,000 people in Uvira. The main water intake for the REGIDESO water supply network, located on the Mulongwe river, was entirely swept away, causing an immediate water supply service interruption affecting the entire network. The project construction team was mobilised to conduct a damage and needs assessment; this informed the deployment of a rapid humanitarian response by OXFAM International and other agencies, and project materials available on site were used to set up a provisional, replacement structure for the intake. This allowed a partial resumption of the water supply service six weeks later, on 29th May 2020, albeit with a reduced production capacity. Network extension plans had to be revised as some pipes laid in preparation for new connections were buried under up to 4 metres of sediment. The construction of a new, raw water intake was integrated into new project components to be implemented at a later stage. The research team remotely coordinated a survey of 148 households in July 2020 in order to document the flooding effects on domestic practices related to water and sanitation; survey results suggest that, while about half of respondents (44.6%) had to change their main water source after the flooding, the situation was ‘back to normal’ after three months, with only 9% of respondents using emergency water supply (bladders) [30].
The security situation in Uvira remained of concern throughout the project as the South Kivu province has experienced protracted conflict since the Congo wars (1996–1997, 1998–2003) and the United Nations Organization Stabilization Mission in the Democratic Republic of the Congo (MONUSCO) has been present in and around Uvira in the past decades. Prevention measures (e.g. curfew) were adopted but despite periods of heightened security alert no major incident affected the project. The partnership with Oxfam International as a humanitarian organisation providing logistical, operational, administrative and security support to the works supervision and research teams proved adequate, including during a necessary emergency evacuation of staff in September 2017.
Contextual elements affecting the water supply service.
The power supply for Uvira primarily depends on the hydropower plant Ruzizi I located near Bukavu, approximately 120 km north of Uvira. Electricity is transmitted via high-voltage lines going through Rwanda and Burundi, and supply interruptions are frequent. According to REGIDESO data, pumps were operated 12.3 hours per day on average between 2017 and 2021 (Fig 4). Informal reports suggest that the electricity supplier may have placed restrictions on the use of electricity for pumping since mid-2020, allowing it only three days per week [31]. The power supply issues have been known since before the project began and a feasibility study for a small hydropower plant on the Mulongwe river was conducted in 2013. This study concluded that the cost would be high relative to the marginal production potential [32] and this plant was not constructed. Likewise, solar options were studied but deemed unsuitable due to a lack of available land, high costs, and technical uncertainties. Although the increase in storage capacity planned with the new 2,000-m3 tank would improve the system resilience to electricity interruptions, the lack of a reliable power supply remains one of the primary reasons for the intermittency of the water supply service and prevents the full exploitation of infrastructure improvements realised as part of the project.
The rise of Lake Tanganyika is a cyclic phenomenon expected every 50–60 years, with effects exacerbated by changes in precipitation patterns linked to climate change [33]. As a result, in Uvira, some lakeside areas are now uninhabitable as they have remained under water since 2020, and the housing has been abandoned. Works to install new water supply connections in these areas had to be cancelled and pipes that had been laid in preparation for the installation of new connections were cut from the rest of the network.
Contextual element affecting the impact evaluation.
A mass OCV campaign was deployed in Uvira in two rounds, from 29th July to 5th August and from 1st to 8th October 2020. The licensed oral vaccine Euvichol-Plus® was offered to all individuals aged one year or older by mobile teams and at fixed vaccination points. The administrative coverage reported by the Ministry of Health was 96.5% for at least one vaccine dose and 93.9% for two doses (full vaccination). Coverage estimates based on a representative household survey conducted in August 2021, approximately one year after the first round of vaccination, were substantially lower: 54.9% (95% CI: 50.5–59.2%) for at least one dose and 23.1% (95% CI: 19.7–26.9%) for two doses [34]. The OCV campaign was expected to decrease the cholera burden in the Uvira Health Zone, without affecting other diarrhoeal diseases; analyses are ongoing as part of another research project led by Johns Hopkins University in order to assess the impacts of mass vaccination in Uvira.
Process evaluation domain #2: Implementation
Overall, considering works completed until December 2021, the water supply service in Uvira did not improve as expected as a result of the project, as shown in a previously published, detailed analysis of REGIDESO operational and billing records [35]. The estimated quantity of water available per user per month even decreased between 2017 and 2021, and it remained under 20 L/cap/day during over 95% of the observation period. This may be due to the fact that the number of connections to the network increased, while the water production capacity remained limited. Reasons for this gap include: (i) the reliance on a provisional water intake installed post-flooding, which reduced the volume of incoming water at the main treatment plant, (ii) delays in construction of the new tank and issues with the waterproofing coating quality preventing its use at full capacity; (iii) recurrent, increasingly frequent electricity interruptions (Fig 4). In addition to the quantity issues, the microbiological quality of drinking water available from REGIDESO service users generally failed to comply with the WHO guideline of <1 colony-forming unit Escherichia coli per 100 ml, with only 5% of 518 drinking water samples collected during a household survey in September 2021 meeting this standard [36].
Furthermore, the project implementation was characterised by multiple delays and the initial project duration (2015–2018) had to be extended several times. While data collection activities related to the impact evaluation of infrastructure improvements stopped at the end of December 2021 due to contractual and resource constraints, a new phase of construction works will likely continue until end of 2023, including replacement of the provisional water intake installed post-flooding. As a consequence of the changing timelines, implementation challenges, and limited resources available, the impact evaluation will not capture any subsequent effects of this new phase of construction work currently underway.
Implementation elements affecting the water supply service.
It was initially foreseen that REGIDESO, as works coordinator, would supervise all construction works and Veolia Foundation would intervene as a technical advisor. However, REGIDESO staff were mobilised late and intermittently to support supervision of the construction works in Uvira. The Veolia Foundation was asked to fill gaps in following up with construction companies, which helped advance works while possibly decreasing the sense of project ownership by REGIDESO.
The operation of new infrastructure remained suboptimal: for instance, anecdotal observations by the research and construction supervision teams suggest that even when an adequate electricity supply was available, the pumps were sometimes not operated to produce water for distribution. Although technical training across REGIDESO staff was not systematically assessed as part of this study, a possible explanation is that REGIDESO human resources in Uvira were limited (e.g. no systematic replacement of retired staff) and technical competences within the team were concentrated in a few experienced individuals rather than institutionalised. Trainings and organisational development support provided as part of the project may have been insufficient. In addition, the motivation of staff to improve performance may have been limited by low wages and by a lack of awareness of how critical their role is for public health and quality of life.
The management of new community taps proved difficult. Three committees were formed with support from an NGO with experience in similar projects in order to manage and operate community taps in the south, centre, and north of Uvira. However, given that most taps never became fully operational, a financially viable operation was not possible. In addition, the common practice of sharing taps between neighbours created informal competition with community taps which was not sufficiently considered in the design of the management system for community taps.
Implementation elements affecting the impact evaluation.
While operational records suggest an increase in chlorine concentrations measured at the outlet of the water treatment plant between 2016 and 2017, with a majority of values between 0.5 and 1.0 mg/L from 2017 onwards, the research team was not able to ascertain the sampling and analytical procedures for chlorine testing and systematic water quality testing was not planned as part of the infrastructure improvement programme. Water quality was thus only assessed based on sampling performed at the household level during cross-sectional surveys conducted by the research team.
Cooperation between the Uvira Health Zone and the research team was crucial for the implementation of research activities, which relied on nurses, medical doctors, a laboratory technician, and cholera Focal Point running the cholera treatment structures. The research team was allowed to train Health Zone staff on specific data collection procedures which were integrated into routine healthcare provision. The individuals responsible for key activities (i.e. laboratory technician, cholera Focal Point) remained in place throughout the entire duration of the research project and many of the routine surveillance procedures established remain in place now.
Data collection related to the water supply service was primarily retrospective, through consultation of REGIDESO operational and billing records during field missions of the research team. As a result, data quality control was limited and high level of uncertainty persisted for some parameters of interest such as the volume of water produced (given the reliance on hours of pumps operation in the absence of a water meter at the outlet of the treatment station).
The research project funded a doctoral thesis, several graduate student projects, and generated publications as well as conference presentations–all led by individuals based in Europe. However, no formal partnership was established with academic institutions from DRC or the region. This was a missed opportunity to leverage local research expertise in order to improve research outcomes, including uptake of results.
Process evaluation domain #3: Population response
Population response elements affecting the water supply service.
The demand from users for new connections to the water supply network was initially lower than expected, and variable. For example, in the first cluster where works were carried out (#10), the installation of 25 new taps was envisioned but only eight users filled out a request for a new tap within four weeks after the start of the works, which was the period allocated for the installation of new taps in each cluster according to the trial protocol. Full flexibility was therefore given to the construction team to adapt the number of new and rehabilitated connections by cluster. In the end, rates of completion varied from almost seven times (in cluster #9) to less than 2% (in clusters #14, 16) of the initially planned number of new connections (Table 2). We identified several potential reasons for this. First, repeated delays in the installation of new connections, assessed for the first time in 2013 and initiated in 2019, may have negatively affected trust and willingness to pay for a new connection. Second, known, frequent water supply interruptions and the absence of service improvements (e.g. in areas like cluster 14, where pressure was low) may have discouraged potential users; temporary service improvements apparently led to spikes in the number of requested new connections. Third, the price to pay for a new connection and for the piped water supply service may have been too high. The subsidised connection fee, initially set at US$ 50 (compared to an estimated, monthly income of US$ 88 per household [37]), increased up to US$140 due to the addition of taxes by different administrative units of the town of Uvira. And data from the September 2021 survey (n = 265) suggest that piped water supplied by REGIDESO was approximately five times more expensive than water purchased from other vendors. Having insufficient financial resources was the primary reason invoked by survey participants for not being connected to the water supply network, followed by the (technical) impossibility to connect their household due to its location. Fourth, the promotion and information on the possibility to have a new connection installed may have been insufficient. Last but not least, the widespread habit of sharing taps between neighbours, reported by almost half of survey participants in September 2021, may have discouraged the installation of individual taps.
Population response elements affecting the impact evaluation.
As a consequence of the variety of healthcare facilities used by the population, the clinical surveillance system set up in collaboration with the Uvira Health Zone could only capture a fraction of diarrhoeal disease and cholera cases. During the survey conducted in September 2021, the majority of respondents indicated pharmacies (70%), private clinics (10%) or traditional healers (6%) as their first go-to place for seeking diarrhoeal disease care. The cholera treatment structures monitoring cases for the impact evaluation were mentioned by only 5% of participants, suggesting that clinical surveillance only captures the most severe and acute cases of diarrhoeal disease.
Discussion
The structured evaluation of the context, implementation, and population response in relation to assumptions outlined in the project’s theory of change allows the identification of key learnings from a complex intervention, designed to improve the piped water supply infrastructure in an equally complex environment with a high burden of diarrhoeal diseases.
The partnership established between the project and research teams, government partners, and an international NGO with capacity to support activities in a humanitarian setting appeared highly relevant
It did not only provide a framework for the collaboration–including access to relevant data–but also enabled the provision of adequate logistical and security support for the implementation of construction and research activities. Furthermore, it facilitated a rapid response to the extreme flooding events that affected Uvira in April 2020. Flexibility in research and implementation contracts (and protocols) was necessary considering the multiple delays and challenges faced in the project. Substantial time and resource contingencies could have been planned upfront considering the complexity of the intervention in a fragile setting.
Cross-sectoral planning and coordination (considering water, energy, land use planning) are critical for the realisation of infrastructure projects and exploitation of their full potential
This was highlighted by the fact that power supply interruptions remained frequent and were the main cause of water supply intermittence in Uvira. Delays due to difficulties in obtaining construction permits and the need to mobilise additional resources for anti-erosion control measures also illustrate the relevance of cross-sectoral coordination to enhance implementation efficiency. Besides cross-sectoral coordination issues, anecdotal evidence revealed an underestimation of needs in terms of training, resources–including staffing–and institutional support to REGIDESO for the operation of the extended, modernised water infrastructure built in Uvira. This may not only have threatened the sustainability of realised investments but was also likely to have been a reason why infrastructure improvements did not result in the expected level of service improvements.
Affordability and financial sustainability should be carefully considered in the design of infrastructure programmes in low-resource settings
Even though subsidised, the price for new network connections was high compared to income levels and increased to 1.6 time the estimated average monthly household income due to various added taxes. This reflects insufficient cross-sectoral coordination among stakeholders in a context of weak public institutions and regulation. and cost is likely to have been a significant barrier for many potential new users of the water supply service. In addition, although invoices for the water supply service were in most cases less than 5% of the estimated average monthly household income, survey results suggest that households using the REGIDESO service paid approximately five times more than those using other water providers (e.g. neighbour, vendor, village chief). Despite this difference, it is unclear whether water tariffs set at the national level and effective payments by users to REGIDESO allowed sufficient revenue to ensure a sustainable operation and maintenance of the new, extended infrastructure in Uvira.
Informal practices such as the widespread sharing of taps between neighbours may impact the use and management of the water supply service
In Uvira, this is likely to have reduced the demand for new connections to the water supply network and for the service provided through community taps built as part of the infrastructure improvement programme. A possibility might have been to further investigate sharing practices in preliminary phases of the programme so that, if community taps were deemed appropriate, governance structures could be established based on existing mechanisms and experiences related to selling tap water.
A locally identified contextual factor of increasing global relevance is the influence of climate change
Two phenomena in particular that are plausibly linked to climate change—the extreme flooding events from April 2020 and the sustained rise of the lake level from 2020 onwards–illustrate how water supply infrastructure and services will come under increased stress. Although the availability of hydrometeorological data remains limited, the proactive, systematic identification of climate-related risks, vulnerable areas, and potential mitigation measures would be useful and may, to some extent, be supported through analyses of satellite imagery [38].
Regarding the research project, the pragmatic combination of multiple approaches, as outlined in the published protocol, was adapted to the evaluation of a complex infrastructure intervention with multiple implementation challenges [24]
While local staff from the Uvira Health Zone were heavily involved in research activities and all research activities were approved by the ethics committee at the School of Public Health of the University of Kinshasa, we consider the absence of partnership with a research or academic institution in DRC as a major weakness of the project structure. New research projects building on this study that have been initiated in Uvira have demonstrated the value of collaborating with Congolese researchers and their institutions in terms of both scientific and contextual knowledge sharing. Despite this critical shortcoming, engagement of an international research team from the time of project inception allowed for a rigorous documentation and evaluation of the programme, producing valuable evidence to inform cholera and diarrhoeal disease prevention strategies.
The work presented herein has several limitations, including a heavy reliance on retrospective data collection and secondary analysis of documents and data from the main study. Population response elements were primarily assessed through repeated cross-sectional household surveys, which may not have captured seasonal variability.
Conclusion
Our process evaluation of a large-scale water supply infrastructure improvement programme in a complex emergency highlighted the importance of five elements for its implementation, namely: (i) multistakeholder partnerships, including with researchers, (ii) cross-sectoral planning, (iii) affordability issues, (iv) informal practices (i.e. tap sharing), and (v) effects of climate change. We recommend for development agencies to more systematically pursue such models of early engagement with researchers, to actively support and promote partnerships with research institutes in countries of implementation, and to consider adopting pragmatic evaluation approaches where necessary for the generation of good-quality, scientific evidence from complex programmes and settings in order to inform and enhance the effectiveness of future infrastructure investments contributing to cholera prevention efforts and to progress towards SDG6.
Supporting information
S1 File. Alternative language abstract: French translation of the abstract.
https://doi.org/10.1371/journal.pwat.0000185.s001
(DOCX)
S1 Table. Theory of change assumptions, related evaluation questions and sources of information across the three evaluation domains (context, implementation, population response).
https://doi.org/10.1371/journal.pwat.0000185.s002
(DOCX)
S1 Checklist. Inclusion checklist: PLOS checklist on inclusivity in global research.
https://doi.org/10.1371/journal.pwat.0000185.s003
(DOCX)
Acknowledgments
We would like to thank the Uvira Health Zone and OXFAM International in DRC for their support to this project. We are grateful to the French Agency for Development and the Veolia Foundation, AGC, Arab Contractors, REGIDESO, and the teams of construction workers, all of whom worked and are working tirelessly to improve the Uvira water supply network while patiently helping document the situation. We also would like to express a sincere “thank you” to survey participants who welcomed our teams in their homes and answered our questions.
Lastly, and most importantly, we would like to acknowledge the contributions of the late Dr Jeroen Ensink. Dr Ensink was instrumental in the conception of this study and it would not have been possible without him. He remains very much in our minds as we conclude this study and is still sorely missed as a colleague and as a friend.
References
- 1.
Hutton G, Varughese M. The Costs of Meeting the 2030 Sustainable Development Goal Targets on Drinking Water, Sanitation, and Hygiene. World Bank, Washington, DC; 2016.
- 2.
Bridging infrastructure gaps: Has the world made progress? | McKinsey. [cited 2 Apr 2023]. https://www.mckinsey.com/capabilities/operations/our-insights/bridging-infrastructure-gaps-has-the-world-made-progress
- 3. Wolf J, Johnston RB, Ambelu A, Arnold BF, Bain R, Brauer M, et al. Burden of disease attributable to unsafe drinking water, sanitation, and hygiene in domestic settings: a global analysis for selected adverse health outcomes. The Lancet. 2023;401: 2060–2071. pmid:37290458
- 4. Ali M, Nelson AR, Lopez AL, Sack DA. Updated Global Burden of Cholera in Endemic Countries. PLoS Negl Trop Dis. 2015;9. pmid:26043000
- 5.
World Health Organisation (WHO). Cholera, 2022. Weekly epidemiological record N°38. 2023. https://iris.who.int/bitstream/handle/10665/372986/WER9838-eng-fre.pdf?sequence=1&isAllowed=y
- 6. Lessler J, Moore SM, Luquero FJ, McKay HS, Grais R, Henkens M, et al. Mapping the burden of cholera in sub-Saharan Africa and implications for control: an analysis of data across geographical scales. The Lancet. 2018;391: 1908–1915. pmid:29502905
- 7. D’Mello-Guyett L, Gallandat K, Van den Bergh R, Taylor D, Bulit G, Legros D, et al. Control and Prevention of Cholera with Water, Sanitation and Hygiene Interventions: a Systematic Review of Current International Cholera Guidelines. PLOS ONE. 2020 Jan 8;15(1):e0226549.
- 8. Legros D. Global Cholera Epidemiology: Opportunities to Reduce the Burden of Cholera by 2030. J Infect Dis. 2018;218: S137–S140. pmid:30184102
- 9. Sikder M, Deshpande A, Hegde ST, Malembaka EB, Gallandat K, Reiner RC, et al. Water, Sanitation, and Cholera in Sub-Saharan Africa. Environ Sci Technol. 2023;57: 10185–10192. pmid:37409942
- 10.
Global Task Force on Cholera Control (GTFCC). Ending Cholera: A Global Roadmap to 2030. 2017. https://www.gtfcc.org/about-cholera/roadmap-2030/
- 11. Wolf J, Hubbard S, Brauer M, Ambelu A, Arnold BF, Bain R, et al. Effectiveness of interventions to improve drinking water, sanitation, and handwashing with soap on risk of diarrhoeal disease in children in low-income and middle-income settings: a systematic review and meta-analysis. Lancet. 2022;400: 48–59. pmid:35780792
- 12.
Joint Monitoring Programme (JMP). JMP Methodology—2017 update & SDG baseline. 2018. https://washdata.org/monitoring/methods
- 13. Kiama C, Okunga E, Muange A, Marwanga D, Langat D, Kuria F, et al. Mapping of cholera hotspots in Kenya using epidemiologic and water, sanitation, and hygiene (WASH) indicators as part of Kenya’s new 2022–2030 cholera elimination plan. PLOS Neglected Tropical Diseases. 2023;17: e0011166. pmid:36930650
- 14. Bwire G, Mwesawina M, Baluku Y, Kanyanda SSE, Orach CG. Cross-Border Cholera Outbreaks in Sub-Saharan Africa, the Mystery behind the Silent Illness: What Needs to Be Done? PLoS One. 2016;11: e0156674. pmid:27258124
- 15. Moore GF, Audrey S, Barker M, Bond L, Bonell C, Hardeman W, et al. Process evaluation of complex interventions: Medical Research Council guidance. BMJ. 2015;350: h1258–h1258. pmid:25791983
- 16. Pfadenhauer LM, Gerhardus A, Mozygemba K, Lysdahl KB, Booth A, Hofmann B, et al. Making sense of complexity in context and implementation: the Context and Implementation of Complex Interventions (CICI) framework. Implementation Science. 2017;12: 21. pmid:28202031
- 17.
Gallandat K, MacDougall A, Jeandron A, Saidi JM, Rumedeka BB, Malembaka EB, et al. Improved water supply infrastructure to reduce acute diarrhoeal diseases and cholera in Uvira, Democratic Republic of the Congo: results and lessons learned from a pragmatic trial. OSF Preprints; 2023.
- 18.
DRC: A Mapping of Non-International Armed Conflicts in Kivu, Kasai and Ituri | Rulac. [cited 15 Jul 2023]. http://www.rulac.org/news/democratic-republic-of-the-congo-a-mapping-of-non-international-armed-confl
- 19.
Internal Displacement Monitoring Centre. Internal Displacement—Country Profile—Democratic Republic of the Congo. 2019. https://www.internal-displacement.org/countries/democratic-republic-of-the-congo
- 20.
Comité National d’Actions de l’Eau, de l’Hygièine et de l’Assainissement (CNAEHA). Cartographie du choléra en RDC: voies probables de propagation et typologie. 2017. https://www.humanitarianresponse.info/sites/www.humanitarianresponse.info/files/documents/files/wash_analyse_cholera_sept_2017.pdf
- 21. Jeandron A, Cumming O, Rumedeka BB, Saidi JM, Cousens S. Confirmation of cholera by rapid diagnostic test amongst patients admitted to the cholera treatment centre in Uvira, Democratic Republic of the Congo. Azman AS, editor. PLOS ONE. 2018;13: e0201306. pmid:30067824
- 22.
Agence Française de Développement (AFD). Note de communication publique d’opération. République Démocratique du Congo. Projet Lutte contre le choléra à Uvira par l’amélioration de l’accès à l’eau potable et à l’hygiène. 2014.
- 23. Jeandron A. Tap water access and its relationship with cholera and other diarrhoeal diseases in an urban, cholera-endemic setting in the Democratic Republic of the Congo. London School of Hygiene and Tropical Medicine; 2020.
- 24. Gallandat K, Jeandron A, Ross I, Mufitini Saidi J, Bashige Rumedeka B, Lumami Kapepula LK, et al. The impact of improved water supply on cholera and diarrhoeal diseases in Uvira, Democratic Republic of the Congo: a protocol for a pragmatic stepped-wedge cluster randomised trial and economic evaluation. 2021. pmid:34154636
- 25. Murdoch J. Process evaluation for complex interventions in health services research: analysing context, text trajectories and disruptions. BMC Health Services Research. 2016;16: 407. pmid:27538946
- 26. De Silva MJ, Breuer E, Lee L, Asher L, Chowdhary N, Lund C, et al. Theory of Change: a theory-driven approach to enhance the Medical Research Council’s framework for complex interventions. Trials. 2014;15: 267. pmid:24996765
- 27. D’Mello-Guyett L, Greenland K, Bonneville S, D’hondt R, Mashako M, Gorski A, et al. Distribution of hygiene kits during a cholera outbreak in Kasaï-Oriental, Democratic Republic of Congo: a process evaluation. Conflict and Health. 2020;14: 51. pmid:32760439
- 28. Bick S, Buxton H, Chase RP, Ross I, Adriano Z, Capone D, et al. Using path analysis to test theory of change: a quantitative process evaluation of the MapSan trial. BMC Public Health. 2021;21: 1411. pmid:34271913
- 29. Jeandron A, Saidi JM, Kapama A, Burhole M, Birembano F, Vandevelde T, et al. Water Supply Interruptions and Suspected Cholera Incidence: A Time-Series Regression in the Democratic Republic of the Congo. Brocklehurst C, editor. PLOS Medicine. 2015;12: e1001893. pmid:26506001
- 30.
Gallandat K, Saidi J, Shetebo J, Azman A, Cumming O. Enquête d’évaluation des changements d’accès aux services d’eau, hygiène et assainissement suite aux inondations extrêmes d’avrili 2020 à Uvira, Sud-Kivu, République Démocratique du Congo. (Report available from the authors.). 2020.
- 31.
Fondation Veolia. Projet de lutte contre le choléra à Uvira par l’amélioration de l’accès à l’eau potable et à l’hygiène. Rapport trimestriel—Assistance à maîtrise d’ouvrage et assistance à maîtrise d’oeuvre. 2022.
- 32.
SHER Ingénieurs-conseils s.a., Mhylab. Microcentrale hydroélectrique sur la Mulongwe, Uvira, République Démocratique du Congo—Etude de faisabilité. Fondation Veolia; 2015.
- 33.
La montée des eaux du lac Tanganyika provoque des besoins humanitaires urgents au Burundi. In: ONU Info [Internet]. 19 Jul 2021 [cited 6 Jun 2022]. https://news.un.org/fr/story/2021/07/1100332
- 34.
Bwenge Malembaka E, Jones F, Dent Hulse J, Musole Bugeme P, Saidi Mufitini J, Bashige Rumedeka B, et al. Campagnes de vaccination orale de masse contre le choléra dans le contexte de la pandémie de la COVID-19 à Uvira, République Démocratique du Congo: enquête de couverture 11 mois après la vaccination. (Report available from the authors.). 2021 Nov.
- 35. Gaiffe M, Dross C, Bwenge Malembaka E, Ross I, Cumming O, Gallandat K. A fuzzy inference-based index for piped water supply service quality in a complex, low-income urban setting. Water Research. 2023;243: 120316. pmid:37453403
- 36. Gallandat K, Hutchins C, Ross I, Bwenge Malembaka E, Lumami Kapepula V, Musole Bugeme P, et al. Evaluation de l’impact sur les maladies diarrhéiques de l’amélioration de l’approvisionnement en eau de la ville d’Uvira, République Démocratique du Congo (RDC). Enquête auprès des ménages sur les pratiques domestiques liées à l’eau. 2022.
- 37.
AFD, Nodalis. Survey of household income in Bukavu. 2021 May.
- 38. Skakun S, Kussul N, Shelestov A, Kussul O. Flood Hazard and Flood Risk Assessment Using a Time Series of Satellite Images: A Case Study in Namibia. Risk Analysis. 2014;34: 1521–1537. pmid:24372226