Objective: The organizational structure of health care facilities has been shown to affect outcome in critically ill patients. We evaluated the association between structures, treatments and outcomes in a large cohort of critically ill patients.
Design: Prospective multicentre cohort study.
Patients and setting: A total of 26,186 patients consecutively admitted to 31 intensive care units (ICUs) in Austria from January 1998 through December 2000.
Measurements and results: The ICUs were divided into three groups according to the size and function of the hospital: community hospitals and specialized trauma centers (group A); central referral hospitals (group B); and teaching hospitals (group C). Group C patients exhibited a significantly higher risk-adjusted mortality (O/E ratio). Although severity of illness at admission in groups B and C was similar, group C patients received significantly more invasive diagnostic and therapeutic interventions throughout their ICU stay: For 7 of 10 invasive interventions identified, odds ratios for group C vs group B patients were significantly increased, even after adjustment for age, gender, severity of illness and reason for admission (odds ratios 1.2-13.1; all 95% CIs >1). Risk-adjusted multivariate analysis confirmed that six of these invasive interventions were independently associated with mortality. Furthermore, nurse-to-patient ratios did not differ between groups, leading to a significantly increased nursing workload in group C ICUs.
Conclusions: Several invasive interventions were independently associated with increased mortality. Our results provide strong evidence that this association was responsible in part for the increased risk-adjusted mortality in group C patients.