Impact of intensive care unit physician on care processes of patients with severe sepsis in teaching hospitals

Objective: The purpose of the study was to investigate associations among intensive care unit (ICU) staffing and care processes in patients with severe sepsis. Design: An observational multicenter cross-sectional study performed from October 2007 to March 2008. Setting: Forty-nine teaching hospitals in Japan.  Participants: Patients (n=576) with severe sepsis identified using ICD-10 codes from administrative data. Main outcome measures: Care processes including mechanical ventilation, dialysis, enteral feeding, parentetal nutrition, and antibiotic empirical therapy which were available in administrative data. Results: ICUs were classified as high- or low-intensity based on policies regarding the responsibilities of intensivists. There were no differences in baseline patient characteristics between the ICU groups. In the high-intensity group, ICU stay for survivors was about two days shorter and hospital stay was significantly shorter by three days. Majority of patients had high rates of enteral feeding; however, the high-intensity group had significantly earlier initiation of enteral feeding and a significantly shorter duration of mechanical ventilation. A shorter duration of mechanical ventilation was significantly associated with the ICU structure. Conclusions: The results showed an association between ICU physician and processes of intensive care, and high-intensity ICU was aggressive in mechanical ventilation in patients with severe sepsis.