We evaluate the effect of a preschool health intervention aimed at reducing infections through improved hygiene practices and training of the staff. The large-scale design enables us to study heterogeneous effects with respect to several child and preschool characteristics that are assumed to be associated with the initial level of absence, and thereby the potential magnitude of the effect. We find no reduction, on average, in children's illness-related absence. This conclusion holds across different age groups of preschool children, and for preschools with varying levels of absence before the intervention. However, we do find some evidence for a decline in illness-related absence among children enrolled in preschool sections with relatively few children. We find some empirical support for that the intervention consisted of components with opposing effects on illness-related absence, which may explain the zero average effect.