A cohort of 4515 surgical patients in ten selected intervention groups was followed. Three hundred and seventeen developed postoperative wound infections, and 291 of these cases were matched 1:1 to controls by operation, sex and age. In comparison to the controls the cases stayed longer in hospital after the intervention and had more contact after discharge with the social security system. Using data from a national sentinel reference database of the incidence of postoperative wound infections, and using national activity data, we established an empirical cost model based on the estimated marginal costs of hospital resources and social sick pay. It showed that the hospital resources spent on the ten groups, which represent half of the postoperative wound infections in Denmark, amounted to approximately 0.5% of the annual national hospital budget. This stratified model creates a better basis for selecting groups of operations which need priority in terms of preventive measures.