Abstract In Belgium, it is mandatory to report Shiga toxin-producing Escherichia coli (STEC) infections to the health inspection authorities. To facilitate the decision making regarding infection control measures, information about the risk factors for the development of the haemolytic uremic syndrome (HUS) can be helpful. We performed statistical analyses on a dataset of 411 Belgian STEC strains. Demographic and clinical patient characteristics as well as phenotypical and genotypical STEC strain characteristics were taken into account. Multivariate logistic regression models indicated that age categories ⩽5, 6–12 and ⩾75; the stx2 gene; and the eae gene were significant HUS development risk determinants. The stx2a subtype had the highest risk (OR 29.6, 95% CI 7.0–125.1), while all stx1 subtypes encompassed a significant lower risk (OR 0.3, 95% CI 0.1–0.5). Presence of the stx1 gene without stx2 encompassed a lower risk than the combined presence of stx1 and stx2 , or stx2 solely. Based on these results, we propose a new virulence typing algorithm that will enable the National Reference Centre to provide the physicians and health inspection authorities with a risk classification for the development of HUS. We believe this will contribute to a more efficient STEC infection control management in Belgium.