Competing risk factors associated with nosocomial infection in two university hospitals

An historical cohort study of risk factors associated with nosocomial infection was conducted in two university hospitals on the East and West Coast of the USA. The purpose of the study was to estimate the relative risk ratios (RR) associated with antibiotic therapy, instrumentation, surgical operations and age with nosocomial infection, and to establish the hierarchical relationship of these factors to each other. A high risk cohort of long-staying patients (average length of stay = 30 days), was selected and each patient's chart was analysed retrospectively for infection occurrence and for risk factors. There was no significant difference between hospitals in rates of nosocomial infections, diagnostic categories and utilization of risk factors. Categorical linear regression analysis showed that a four factor model consisting of antibiotic therapy, instrumentation, surgical operations and age accounted for 95 per cent of the variation in nosocomial infection rates. Length of stay was treated as a co-dependent variable. Multivariable stratification analysis yielded aetiologic fractions of 63 per cent attributable to antibiotic therapy, 26 per cent to surgical operations, and 13 per cent to instrumentation. About half the antibiotic therapy was given on admission or at least 4 days before the first hospital-acquired infection. These data support careful assessment of prophylactic antibiotic therapy and infection control policies to reduce risk of nosocomial infection in university hospitals.