Classification and ranking of shigatoxin-producing Escherichia coli (STEC) genotypes detected in food based on potential public health impact using clinical data

Risk classification and management of shigatoxin-producing E. coli (STEC) isolated from food has been hampered by gaps in knowledge about the properties that determine the extent to which different subtypes of STEC can cause severe disease. Data on the proportion of infected human cases being affected by severe illness enables an evaluation of existing approaches for classifying STEC strains and the development of a new public health based approach. Evaluations show that existing approaches do not unequivocally classify different STEC variants according to their ability to cause severe disease. A new approach for ranking of STEC genotypes, combining the estimated probability of the strain to cause severe illness with the public health burden associated with the illness in terms of DALY per case, address these limitations. The result is a list of STEC genotypes in descending order of potential public health burden per case. The approach is risk based in considering the probability and consequences following infection (severe illness), and can support transparent risk management. This is illustrated by, arbitrarily, separating the ranked list of genotypes into classes based on the potential public health burden, and by characterising collections of strains isolated from different foods into different classes. Further, the classification of food samples as satisfactory or not based on the cost in terms of proportion of food being rejected and the benefit in terms of the proportion of strains causing severe illness (HUS) that are being captured is demonstrated using this approach.