Circuit Rider post-construction support: improvements in domestic water quality and system sustainability in El Salvador

Small piped water supply systems are often unable to provide reliable, microbiologically safe, and sustainable service over time, and this has direct impacts on public health. Circuit Rider (CR) post-construction support (PCS) addresses this through the provision of technical, financial, and operational assistance to these systems. CRPCS operates in low and high-income countries; yet, no rigorous studies of CRPCS exist. We measured the impact of CRPCS on ‘water quality’ and ‘sustainability’ indicators (technical and administrative capacity, and water supply protection) in El Salvador. In this field-based study, a case-control design was utilized in 60 randomly selected case (28 CR) and comparable control (32 noCR) communities. Microbiological water quality tests and pretested structured key-informant interviews were conducted. The operational costs of CRPCS were also assessed. Data were compared using parametric and non-parametric statistical methods. We found communities with CRPCS had significantly lower microbiological water contamination, better disinfection rates, higher water fee payment rates, greater transparency (measured by auditable banking records), greater rates of household metering, and higher spending for repairs and water treatment than comparable control communities. CRPCS is also a low-cost (<$1 per household/year in El Salvador) drinking water intervention. doi: 10.2166/washdev.2014.136 om https://iwaponline.com/washdev/article-pdf/4/3/460/384962/460.pdf 2020 Georgia L. Kayser (corresponding author) The Department of Environmental Science and Engineering and Public Policy, The Water Institute, Gillings School of Global Public Health CB#7431, University of North Carolina at Chapel Hill, NC 27599-7431, USA E-mail: gkayser@unc.edu Georgia L. Kayser William Moomaw The Center for International Environment and Resource Policy, The Fletcher School of Law and Diplomacy, Tufts University, Medford, MA 02155, USA Jose Miguel Orellana Portillo Asociación Salvadoreña de Sistemas de Agua, 7a Calle Oriente, #28 Bo El Santuario, San Vicente, El Salvador, Central America Jeffrey K. Griffiths The Department of Public Health and Community Medicine, Tufts University School of Medicine, 136 Harrison Avenue, Boston, MA 02111, USA

obstacles. PCS provides technical assistance in operation and maintenance (O&M), and administrative and financial management training to community-run water systems.
A review of the literature reveals that few PCS impact studies exist. A retrospective World Bank study of PCS in Bolivia, Ghana, and Peru, suggests that communities whose system operators attend training workshops, experience better system performance than non-PCS communities (Bakalian & Wakeman ; Whittington et al. ). PCS was associated with improved financial performance and household satisfaction in Peru (Prokopy et al. ). In Bolivia, however, engineering-oriented PCS had no measurable impact on system function or user satisfaction (Davis et al. ). The best configuration for PCS is unknown, as are its impacts upon microbiological water quality and its cost. CRPCS consists of support in four main areas: technical, financial, and administrative management, and water supply protection (see Table 1). CRPCS is provided through trainings, on-call technical support, monthly visits, and capacity building workshops from a government agency or local NGO in El Salvador (Kayser a, b, ).
Monthly visits allow CRPCS technicians to test for chlorine disinfection and microbiological water quality and address community-specific management needs. Workshops provide training in O&M, water treatment operation, budgeting and accounting, water supply protection, and household metering for operators and Village Water Committee (VWC) members. One CR technician visits the same, approximately, 25 communities every month.
CR technicians first examine the DWS and facilitate a needs assessment. This appraisal includes system conditions (from the source through its treatment and distribution), and VWC activities (Kayser a, b, ). These include: the presence of a VWC, VWC responsibilities, operator technical assistance, administration of financial accounting and bookkeeping, household water fees, presence of a VWC bank account for user water fees, presence of water meters, existence of a supply inventory, and plans for maintenance and operation. The technician tests the water for disinfection (residual chlorine) and microbiological quality (Escherichia coli and total coliform bacteria). The assessment informs community-specific PCS trainings. CR In this field-based study, a case-control design was utilized to assess the impact of CRPCS on 'water quality' and households in four provinces of the country. Our objective was to measure the actual effectiveness of CRPCS as implemented, rather than potential impact under ideal conditions. We then contrast the impacts of this water supply intervention and its costs.

Ethics
Free and informed participant consent was obtained. The

Institutional Review Board of the Fletcher School of Tufts
University approved this protocol on June 13, 2007.

Research design
We used a case-control design to assess how CRPCS may impact water quality and sustainability metrics in randomly selected intervention (CR) and control (noCR) communities.
We randomly selected 60 villages (28 CR and 32 noCR) using primary and secondary data. Primary data included lists of DWSs from regional municipal offices, and a roster of CRPCS communities from ASSA. Secondary data included census information and department maps. Intervention communities were randomly selected from the ASSA roster. Controls were selected in two steps: first, selecting for the presence of a piped DWS from the list compiled from regional municipal offices, and then using geographic data to identify communities similarly located to the randomly chosen intervention communities.

Data collection
Field study data collection occurred in February 2009.
The operator and the treasurer or president of each VWC was interviewed using previously piloted structured interviews to identify key sustainability metrics. System sustainability metrics were assessed through interview questions in four categories comprising the Sustainability Index: financial, technical, and administrative management, and water supply protection ( To assess water quality, standard microbiological (E. coli and total coliform) and residual chlorine tests were run on samples obtained from the first and last households on the piped system, which were GPS geocoded.
Chlorine residual assays were done on site using HACH DPD free chlorine reagent powder pillows. If communities had more than one DWS, both were tested. Microbiology tests were performed offsite as outlined below.
Sterile Whirl-Pak ® bags (Nasco, Modesto, CA) were used to collect 100 mL of water, which was coded and placed on ice.

Data analysis
We used t-tests to assess normally distributed data, Mann-Whitney U non-parametric tests to evaluate non-normal data, Chi-square tests for frequency data, and Fisher exact tests when these frequency data were small (<5) (see Tables 3-7) using IBM ® SPSS version 19 software.
The data were also post-hoc re-analyzed without one control community. It had a DWS that was privately run with federal operational funding, unlike all other sampled DWSs that were community run and managed. By chance, this community had been randomly chosen during the selection process. This DWS had the highest number of households served, construction costs, and operating budget. Our results were essentially identical with or without its inclusion. We report the results of analyses with its inclusion.

Comparability of case and control communities
To characterize the comparability of case (CR) and control communities (noCR) and their DWSs, we compared physical factors that could contribute to improved water quality or system sustainability.   More CR villages reported leaky DWS pipes, perhaps indicating enhanced awareness of system needs, given their significantly higher spending on system repairs than noCR communities (p < 0.05) ( Table 5).

Improved financial management in CR communities
CR communities had significantly greater water bill payment rates (p < 0.05), more spending on DWS treatment and repairs, were more knowledgeable about DWS costs (p < 0.05), and were more likely to have water fees deposited  into a bank account than into the hands of a community member (p < 0.05). All but one community charged water service fees. Similar amounts were charged by noCR and CR communities for household water service. Monthly operator wages, and energy costs, did not differ between CR and noCR communities. Financial constraints, such as the energy costs for water pumping, were cited as reasons for reducing household water supply.
DWS investments for O&M such as repairs, operator wages, and water treatment were significantly greater in CR communities than in noCR communities (p < 0.05).
O&M investment was a mean of $509 per month in noCR communities as compared to $1,310 in CR communities.

Much of this difference was related to investments for
repairs and water treatment. VWC debt was also higher in CR communities as a result of loans for investment in operations (Table 6).
Administrative management did not differ between communities CR communities were slightly more likely to have a VWC, have women VWC members, and to pay operators a higher wage; but these results were not statistically significant (Table 7).
CR communities were more likely to conserve water supplies Water meters were more common in CR than noCR communities (p < 0.05). Metered communities typically charged a baseline fee for a basic household water allotment, and an additional fee for water consumed above the baseline. Water supply protection via foresting or fencing, however, was similar in both groups. Only one CR community had a watershed reforestation project (Table 7).

Operating costs of CRPCS
The annual operating cost of CRPCS was less than $1 USD  Perhaps increasing confidence in these data, the DWSs were comparable (Table 3), and no evidence of greater community motivation (such as higher DWS operator salaries or

CONCLUSION
The CR model of PCS, as implemented in El Salvador, is associated with improved community drinking water quality, improved financial management, better technical capacity, and a higher prevalence of household metering.
CR communities had significantly less microbiological water contamination, and invested significantly more on treatment and on repairs than comparable control communities. This suggests better maintenance and operation, and long-term sustainability. These positive outcomes were found by assessing a functioning support program, rather than through a best-scenario efficacy trial, suggesting that these benefits represent what is found during actual implementation. We found CRPCS to lead to significantly less water contamination and better sustainability, at a cost of less than $1 per household per year.

DISCLOSURE
The authors declare no conflict of interest. To assure impartial evaluation, Jose Miguel Orellana Portillo assisted in the design of the CR model in El Salvador, conception of the research idea, and writing about the CR model, but did not contribute to collection and or analysis of these data.