Report No DWI0784

Oct 1996

Microbiological risk assessment models have been used to define targets for the degree of pathogen removal by drinking water treatment (Surface Water Treatment Rule) and to identify the maximum concentrations of pathogens in drinking water (Maximum Contaminant Levels) to ensure acceptable risk from waterborne pathogens to consumers.

In this contract, funded and managed by the Drinking Water Inspectorate, components have been assembled to develop a risk assessment model for Cryptosporidium. in drinking water supplies. Existing risk assessment models developed for Cryptosporidium, Giardia and enteric viruses in drinking water rely on the assumption that pathogen organisms are randomly dispersed within 100 litre or 1,000 litre volumes. One manifestation of this assumption is that current models predict that consumers are either exposed to zero pathogens or to doses of just one pathogen. According to current models the probability of two or more pathogens being present in the same glass of water is negligible Available evidence in the literature shows this assumption is not appropriate for drinking water.

Experiments performed in this contract have demonstrated that aerobic spores, which may serve as surrogates for Cryptosporidium oocysts, are not randomly dispersed in drinking water but are spatially associated to some degree, with some 100-ml subsamples containing considerably higher spore counts than others. It has been possible to accommodate some of the spatial variation by fitting the negative binomial statistical distribution. However, a small number of 100-ml samples contained very high counts which could not be accommodated. Although rare in occurrence, those high count samples could be of public health importance and need to be included in risk assessment models.

Overall, the data and most of the methodology to develop a model for assessing risks from waterborne cryptosporidiosis have been established in this contract. This is the first microbiological risk assessment model to account for some of the clustering of microorganisms within drinking water samples. However, before a Cryptosporidium risk assessment model can be constructed further statistical consideration is needed to:

1. define the effect of drinking water treatment on the dispersion of oocysts,
2. and to accommodate the few samples with very high counts.

From experiments with aerobic spores reported here, the data necessary to achieve the first of these two tasks are available. However, to characterise the rare occurrences of very high counts, much larger numbers of samples will be required.

Copies of this report may be available as an Acrobat pdf download under the 'Find Completed Research' heading on the DWI website.