From data to decisions: understanding information flows within regulatory water quality monitoring programs

Diarrheal diseases from unsafe drinking water cause an estimated 230,000 deaths annually in sub-Saharan Africa, and up to 1.8 billion people globally consume water likely contaminated microbiologically. Most countries require regulatory monitoring of drinking water—operational monitoring by suppliers and surveillance monitoring by public health agencies—to mitigate risks. While many institutions conduct microbiological testing, it is unclear whether the resulting data are effectively used. Prior work has largely focused on the structural aspects of testing programs (extent of testing, low-cost test tools, sampling logistics, and mobile data collection). A systems-level analysis indicates institutional commitment and capacity underpin successful monitoring programs, and that many programs are data-rich but information-poor, failing to link collected data to decision-making. Misalignment between information producers and users often leads to dissatisfaction and non-use of data. Earlier research found local actions in response to contamination often occur when producers and users are the same entity, but evidence was limited on whether results reliably reached senior managers, regulators, or other decision-makers. Given the cost and effort of testing, maximizing cost-effectiveness demands timely and useful data transfer to those managing water safety and enforcing standards. This study assesses formal and informal systems used by institutions with regulatory testing responsibilities in sub-Saharan Africa to organize, analyze, and transmit information, maps information flows, compares practices to policy (with a Kenya case study), and proposes recommendations to improve water quality information flows and support water safety management.

The paper synthesizes prior research showing: (i) assessments of the extent of testing practices by suppliers and agencies; (ii) the development and evaluation of low-cost fecal indicator organism tests; (iii) guidance on sampling frequencies, locations, and logistics; and (iv) the use of mobile phone-based tools for data collection. A systems-level comparative analysis highlighted institutional commitment, leadership, knowledge, and staff retention as critical to high-performing monitoring programs. Long-standing concerns persist that monitoring programs are data-rich but information-poor, with collected data not feeding back into management decisions—an issue well-documented in environmental water monitoring but less so in drinking water systems. Research also indicates frequent misalignment between the goals of information producers and users; users are often not involved in monitoring network design, limiting data uptake. Prior observations in sub-Saharan Africa suggested that many institutions take action on contamination when the same entity both produces and uses the information, but downstream transmission to higher-level decision-makers and regulators is uncertain. Given testing is costly and resource-intensive, there is a need to design monitoring programs that generate decision-relevant information and ensure effective information flows.

Study design: The study was conducted within The Aquaya Institute’s Monitoring for Safe Water (MfSW) program. Partnerships were established with 26 institutions across six countries (Ethiopia, Guinea, Kenya, Senegal, Uganda, and Zambia): 11 water suppliers (national, regional, municipal, and a private operator association) responsible for operational monitoring, and 15 surveillance agencies (one national health ministry, three regional surveillance laboratories, and 11 district health or water offices) responsible for surveillance monitoring. Data collection (2012–2016): From November 2012 to July 2015, the team collected qualitative and quantitative data on microbial water quality monitoring via needs assessments, midterm assessments, and ongoing communications. Methods included semi-structured interviews with laboratory technicians and managers, and observations of sampling, testing, and reporting processes. Institutions conducted microbial water quality testing and submitted monthly electronic results to the research team. MfSW support included start-up funds (equipment and training), monthly payments per completed test (above baseline up to a target), and bonus payments for meeting targets. Institutions were required to submit digital results to receive payments. These incentives likely improved information flows, representing a best-case scenario. Follow-up (2019, Kenya): Four years post-intervention, follow-up research in Kenya covered seven prior MfSW institutions (four water suppliers; three county/sub-county public health offices) and six additional county/national ministries and agencies. Approximately five staff per institution were interviewed (directors, chief public health officers, PHOs responsible for water quality, WASH coordinators, lab managers and assistants). A quantitative assessment of sampling, testing, and reporting actions was also conducted. Data flow diagramming: Using NVivo to code interviews, observations, surveys, and notes, the team constructed Data Flow Diagrams (DFDs) following Kendall and Kendall conventions to map inputs, processes, data stores, and outputs for each of the 26 institutions and two Kenyan regulatory agencies. DFDs captured flows from water sources or distribution systems through internal institutional processes to external recipients (e.g., ministries, regulators, management). Patterns across institutions were analyzed, including external entities reported to and the formats of data stores. Ethical review: The study was deemed exempt under the Common Rule 45 CFR 46.101(b)(2) by the Western Institutional Review Board.

• Structural similarity of information systems: Across 26 institutions, suppliers and surveillance agencies followed similar data flows: selecting sampling locations; collecting and recording field information; field and/or laboratory testing; recording and compiling results; digitization; analysis/summarization; reporting to external entities; and initiating actions in response to contamination.
• Roles and processes: Multiple personnel groups typically handled different steps (sampling, testing, data entry, management, reporting). Kenyan county public health offices commonly involved a PHO (sampling/analysis), an M&E Officer (digitization), and head county PHO (review/submit). Suppliers used lab technicians/assistants for testing and managers for review and reporting.
• Data stores and planning: Most suppliers had written sampling plans with set schedules and repeated locations (Kenya: 4/4; overall: 10/11). Fewer surveillance agencies had written sampling plans (Kenya: 1/3; overall: 7/15), and they rarely used set schedules or repeated locations. Field data were mostly recorded on containers, blank paper, or templates; all institutions transcribed microbial results from paper to digital formats at some point, often driven by the MfSW digital submission requirement.
• Reporting and feedback: All institutions reported to at least one national administrative unit (health ministry, water/environment ministry, regulator, or national boards). Surveillance agencies had more and more varied reporting routes than suppliers, often sending to local government units and stakeholder forums. However, institutions rarely received feedback from external entities (acknowledgements, questions, or formal responses), with feedback primarily occurring internally via upper management review.
• Barriers to efficient data sharing: Inefficient paper-based transfers, limited computer/internet access, and lack of integrated databases hindered data flows. Kenyan surveillance agencies reported DHIS only captures numbers of tests, not results; suppliers report via a separate system (WARIS) to WASREB. Both surveillance and supplier institutions expressed the need for digitization at point of entry and centralized/national databases.
• Actions taken on contamination: Institutions generally acted on positive findings. Suppliers more often verified contamination and mitigated risks; surveillance agencies commonly engaged consumers. Examples included resampling, investigating contamination sources, adjusting chlorine dosing, repairs, flushing lines, closing supplies, consumer engagement, and recommending household water treatment. Reported frequencies included: suppliers verifying contamination (10/11) and mitigating risks (9/11); surveillance agencies engaging consumers (14/15), verifying contamination (4/15), mitigating risks (5/15).
• Kenya policy-practice gaps: WASREB’s WARIS requires quarterly/annual reporting on planned vs conducted tests and compliance with standards, contributing to utility performance scores. The four Kenyan suppliers studied were not compliant with WARIS microbial reporting schedules and faced no penalties beyond lower performance ratings. MoH oversight of county public health is limited under devolution; DHIS does not store water quality results. In practice, revisited surveillance agencies were not reporting water quality data to MoH.
• Post-incentive testing decline: Four years after MfSW ended, most revisited institutions (5/7) were no longer conducting routine microbial testing, primarily due to resource constraints (broken equipment, lack of reagents, transport). Testing, when conducted, was often reactive (complaints/outbreaks) and skewed toward basic physicochemical parameters (pH, turbidity, residual chlorine) rather than microbial indicators.
• Overarching barriers: Limited aggregation and analysis of data (little synthesis into trends), weak enforcement of testing/reporting requirements, limited feedback loops, and fragmented information systems were the key impediments to effective information flows and data use.

The study demonstrates that while institutions possess structured processes for collecting and transmitting water quality data, key weaknesses impede the conversion of data into decision-relevant information. First, reliance on single-sample responses without routine synthesis limits understanding of temporal and spatial trends, reducing value for strategic planning. Building data literacy, routine digitization, and capacity to summarize, analyze, and interpret data (e.g., descriptive statistics, trend graphs) can enhance information usefulness. Second, feedback loops are weak: surveillance agencies disseminate widely but receive little feedback, and there is scant evidence that monitoring results inform iterative redesign of monitoring programs. Strengthening two-way communication at all levels (households, local management, regulators) would help close the information cycle. Third, weak regulatory enforcement diminishes compliance with monitoring frequencies and reporting schedules, reducing data availability for decision-making. Technology solutions (mobile/cloud platforms) are promising but should be matched to context; many testing activities occur at centralized labs where basic computers and spreadsheet software may offer immediate gains. Overall, improvements should focus on fixing gaps within existing systems rather than creating parallel reporting frameworks, with particular emphasis on enforcement, capacity, digitization, and integration with other sectoral information systems.

This study maps and evaluates formal and informal information flows in regulatory drinking water quality monitoring across 26 institutions in six sub-Saharan African countries, revealing broadly similar structures but common bottlenecks: limited data aggregation/analysis, minimal feedback, and weak enforcement of reporting requirements. Case findings from Kenya highlight policy–practice gaps (WARIS and DHIS limitations) and a post-incentive decline in routine microbial testing due to resource constraints. The authors recommend: (1) strengthening enforcement and accountability for testing and reporting; (2) building staff capacity for data management, analysis, and interpretation, potentially including dedicated data roles; (3) prioritizing digitization at point of entry using accessible tools (e.g., spreadsheets) and developing integrated databases that capture actual results, not just counts; and (4) integrating water quality data collection and reporting with other water and sanitation performance information to improve efficiency and decision-making. Future work should evaluate scalable approaches to enforce compliance, build data literacy, and link water quality information systems with broader sectoral data platforms to support adaptive, data-driven water safety management.

• Best-case bias: Information flows were observed during the MfSW program, which provided financial and logistical support and required digital data submission, likely improving testing and reporting relative to routine operations. Follow-up in Kenya showed reduced testing after incentives ceased.
• Mapping completeness: Given the complexity of actors and processes, not all DFDs were verified with institutions, so some details may have been misinterpreted or omitted.
• Scope: The analysis focused mainly on microbial drinking water quality data in built systems, not on water resources, raw water supplies, or treatment process operational data.