Analysis of Cross-Infection Mechanisms in a Hospital System

Abstract

The aim of this work was to examine the relevance of factors related to cross-infection in hospitals, based on a large cross-sectional survey, and to formulate a mathematical model depicting the significance of patient parameters, which will be used to standardise future experimental studies. The analysis showed that age, sex, wound drainage and wound contamination are related to wound infection; and age, sex, antibiotic use and length of hospitalization are related to the nasal carriage of antibiotic-resistant Staphylococci. The significance of the parameters age and length of stay were quantified. Correlation and multiple-regression analyses confirmed the validity of these observations. Models developed were based on the assumption that nasal carriage of antibiotic-resistant Staphylococci is a measure of certain aspects of susceptibility to infection. From multiple-regression analysis, a carriage-rate profile representing the relationship between age, length of hospitalization and nasal carriage rate was derived. That the surface shifts its origin depending on the state of the parameters sex, antibiotic use and environmental (ward) conditions is shown. A model for the relationship between age and nasal carriage was developed from theoretical concepts of ageing and acquisition of resistance applied to the total patient population. A deterministic model based on epidemiological principles was developed to represent the relationship between length of hospitalization and nasal carriage. During the development of the final comprehensive model, the significance of a further variable, the distribution of the number of patients in age and length of stay groups was established. This is a major parameter previously ignored by other researchers and particularly important in comparing infection and nasal carriage rates among different patient populations. The distribution of the number of patients was included as a variable together with age and length of hospitalization in this final model.