Preventing nosocomial infections in the intensive care unit--lessons learned from outcomes research.

Patients in an ICU are at increased risk for a nosocomial infection. Infection control practices to reduce these risks have often been based on scant information. A recent trend to base infection control practices on actual patient outcome data has often provided surprising results. Basic measures such as good handwashing and appropriate patient isolation must be followed. Routine venous catheter placement does not increase the risk of bacteremia, and increases procedure morbidity. The role of different catheter dressings and antibiotic-impregnated catheters in reducing bacteremia is unclear. Nosocomial pneumonias and ventilator-associated pneumonia are common in the ICU. Outcome studies suggest that infrequent changes of ventilatory circuits do not increase the risk of ventilator-associated pneumonia, while allowing substantial cost savings. Manipulation of the pH or flora of the gastrointestinal tract seems to have little influence on patient outcomes, even if there may be a slight reduction in nosocomial pneumonias. Although large randomized trials may be outside the scope of hospital infection control programs and ICUs, any hospital should be able to implement outcomes-based studies of changes in infection control policies and procedures.