Community surveys of self-reported diarrhoea can dramatically overestimatethe size of outbreaks of waterborne cryptosporidiosis

This paper describes a community-based study undertaken to assess the size of a waterborne outbreak of cryptosporidiosis in the North West region of E ngland. The outbreak was linked to a single reservoir in the English Lake D istrict and provided drinking water to over 1.2 million people. There were some 308 laboratory confirmed cases. We conducted a community-based survey for self-reported diarrhoea in four towns within the outbreak area and four control towns. The rate of self-reported diarrhoea was higher in the contr ol towns than in the outbreak towns. It would appear that retrospective com munity-based studies of diarrhoeal disease are subject to recall bias that would overestimate the incidence of illness, especially following reporting in the media. In the light of our findings, we reviewed the study undertak en during the Milwaukee outbreak that produced the estimated size of 405,00 0 cases. It is suggested that the estimate of the size of the Milwaukee out break is severely flawed, and the actual size of this outbreak was between 1 % and 10 k of that claimed.