Audit in infection control--data analysis and infections in the intensive care unit

The importance of high standards of infection control practices in intensive care units is widely accepted. Despite this, few objective measures of the efficacy of such practices have been developed. This hampers clinical audit, which is essential if nosocomial infections are to be kept to a minimum. Small sample size, multiple dependent variables, difficulty in matching populations, and the covert effects of confounding factors and bias, all add to the problems of designing meaningful audit programmes in intensive care settings. Nevertheless, useful analysis can be attempted now that appropriate standards and scoring systems are available. Computer systems can simplify data collection, analysis and reporting but they are no substitute for careful preparation and good statistical technique.