According to a national prevalence study in 2004, one of 14 patients admitted to a Swiss hospital – and even one of 10 in large hospitals – was suffering from a healthcare-associated infection (HAI). In the same year, the annual mortality due to HAI was estimated at 2000 in Switzerland. Hospital size was associated with an increased risk for HAI, but risk factor-adjusted HAI rates were not significantly different and caution must be exercised when attempting to benchmark quality of care using infection rates. Patients in non-acute care sectors at the University of Geneva Hospitals carried a higher burden of HAI than those in acute care. Patients exposed to surgery carry a double burden of HAI compared with those non-exposed; surgical site infection adds to the equal load of non-surgical site infection. All these examples demonstrate that in-depth analysis on the systems level can produce insights that help to design more effective preventive strategies. Additionally, data mining techniques might prove effective to meet the challenge of surveying infection control parameters in real time. If human errors are the cause of many quality flaws, the specificities of human behaviour have to be taken into account when designing infection control procedures. Social interaction and the work environment influence human behaviour. For this reason, we used marketing techniques and human factors engineering to design hand hygiene promotion tools. Together, humans and their environment constitute systems. Systems thinking, a relatively recent science to understand complex systems and the interplay of their components, could help to improve infection control with a technology that produces a better fit between human behaviour and the realities of the work environment. Today, still, at least one in three HAI is preventable!