We agree with Martha Sinclair and Christopher Fairley1 that the existing evidence linking drinking water to gastrointestinal illness in developed countries is not comparable to that of waterborne outbreaks. We also agree with most of their comments about potential biases of time series studies on water turbidity and need for better designs.2 However, we disagree in some other of their comments.

Gastroenteritis caused by chlorine resistant organisms is not so uncommon. In the United States, the estimated incidence of waterborne infection by Cryptosporidium parvum is the second highest pathogen identified, after viruses.3 In England, cryptosporidium infection was identified in 2% of the patients with presumed infective diarrhoea,4 whereas infectious intestinal disease occurs in one in five people each year in England, but only a small proportion of cases is recorded by national laboratory surveillance.5

Secondly, their major criticism to the papers by Schwartzet al 6 7 is about the rather poor correlation between water turbidity and levels of infective agents. If there is a correlation, although poor, the results of this information bias, if any, could be the dilution of the association, instead of a spurious association.

In our opinion, the papers by Schwartz appear as a warning of a new potential public health problem, and although we agree with the final comment of Sinclair and Fairley that “...we risk commiting large expenditures on changes in water treatment technology that may be unnecessary and produce no tangible benefit to public health”, these results suggest that we risk to produce some health damage if we do not improve the technology. Results from well conducted individual studies besides data from adequate surveillance systems will provide a great amount of evidence, but, quite frequently, decisions in public health must be taken without complete knowledge.The ball is in the court of public health decision makers who have to manage weak evidences of a potential association with a precautionary view.