Apart from the United States, which issues regular reports on waterborne outbreaks, few countries have a specific surveillance system in place. In France, surveillance of these outbreaks relies on voluntary reporting by physicians of clusters of acute gastroenteritis (AGE) to health authorities or of non-compliant results from water quality testing. According to these sources, between 4 and 5 outbreaks are recorded annually, but this number is likely significantly underestimated. Each of these outbreaks involves an average of 200 cases. Some are massive, such as the one in Vif, Isère, in 2016, where more than 1,500 cases were reported.

Investigations into these outbreaks typically point to the same underlying causes: rainfall events leading to pollution and/or flooding of the water source, operational incidents (disinfection failure, filtration issues, etc.), distribution incidents (pipe breaks), or backflow of wastewater into the drinking water system.

Water systems at risk are those whose intakes are unprotected or poorly protected against pollution, whose treatment facilities are undersized, whose disinfection systems are absent or inadequate, or where there is no system—or a defective one—to prevent wastewater backflow into the distribution network.
In France, mountainous or volcanic areas where the terrain necessitates a water distribution system consisting of numerous small water sources that are difficult to secure are more frequently confronted with this issue. Areas with highly permeable soils, such as karst regions, are also at greater risk.

No approach currently being developed in France for detecting acute gastroenteritis outbreaks is specific to waterborne transmission. The challenge in optimizing the detection of waterborne outbreaks is to identify clusters of acute gastroenteritis cases that share a common exposure to tap water. This must be done while accounting for the constraints imposed by Health Insurance data, which provide aggregated AGE cases by day and by municipality, and by the boundaries of drinking water networks, which do not directly align with municipal boundaries.

To address this challenge, a decision-making algorithm was developed to define which municipalities should be grouped based on the water distribution units that supply them. A cluster detection method (Kulldorff’s spatio-temporal scan) was then applied to these groupings to identify potential waterborne outbreaks.

A simulation study was conducted to evaluate the performance of this algorithm in detecting clustered cases. First, the baseline incidence of waterborne outbreaks at the municipal level was calculated using data from the French National Health Insurance (background noise). Next, 2,000 outbreaks with profiles consistent with waterborne outbreaks that occurred in France and were described in the literature were generated at the water distribution unit level.

The overall sensitivity of the method is close to 74%. It varies primarily depending on the size of the outbreak (less than 20% for outbreaks involving fewer than 10 cases and greater than 95% for outbreaks of 20 cases or more). The algorithm’s ability to detect a simulated waterborne outbreak is 90.5%. This value is slightly lower for small outbreaks (fewer than 10 cases) and for outbreaks occurring in winter due to higher background levels of acute gastroenteritis during this time of year.

In addition to the simulation study, the detection method was tested in the field as part of a pilot study covering seven departments in seven different regions. The results indicate that the deployment of a retrospective surveillance system for waterborne outbreaks across all French departments is feasible. This system should align with international and national objectives promoted by the WHO, aimed at complementing surveillance based on the quality of distributed water through a quality management system (Drinking Water Management and Safety Plans—PGSSE).
The objectives of the surveillance will be to detect waterborne WEA outbreaks; to conduct investigations into the affected water systems; and to populate a national database with epidemiological and environmental information. Subsequent analyses may be conducted using this database to describe risk factors, assess the epidemiological situation, provide feedback, and inform prevention efforts.
It is estimated that a surveillance system based on the use of Health Insurance data and SISE-Eaux data would increase the number of detected outbreaks by a factor of 100, or approximately 500 outbreaks per year in France.

Implementing this surveillance will require the involvement of a range of partners, each with their own area of expertise: Santé publique France for epidemiological surveillance and the evaluation of prevention measures; regional health authorities (ARS) and operators for environmental investigations and the implementation of management measures; and the Ministry of Health for oversight and any necessary adjustments to regulatory texts.