Methodologic issues in hospital epidemiology. II. Time and accuracy in estimation

Nosocomial infections comprise a variety of diseases that have all of the characteristics and problems traditionally associated with the epidemiology of chronic disease. Each infection is multifactorial in origin, and most of these factors remain either unknown or unquantified. The microbial agent is not the sole determinant of outcome. Time is also an important variable, especially with regard to variation in incubation (latent) period for different infections, and variation in duration of hospitalization (exposure and follow-up). Methods used for the study of chronic diseases may be employed in evaluating some aspects of nosocomial infection. When time is considered as a variable, it is possible to use a common, interconvertible set of definitions of rates that includes measures of both prevalence and incidence of infection. Accuracy in estimation also may be improved by further consideration of the choice of subjects (validity of selection), experimenter effects, misclassification (validity of information), and confounding (validity of comparison.