A multicentre, prospective study to evaluate costs of septic patients in Brazilian intensive care units

BACKGROUND: Sepsis has a high prevalence within intensive care units, with elevated rates of morbidity and mortality, and high costs. Data on sepsis costs are scarce in the literature, and in developing countries such as Brazil these data are largely unavailable. OBJECTIVES: To assess the standard direct costs of sepsis management in Brazilian intensive care units (ICUs) and to disclose factors that could affect those costs. METHODS: This multicentre observational cohort study was conducted in adult septic patients admitted to 21 mixed ICUs of private and public hospitals in Brazil from 1 October 2003 to 30 March 2004. Complete data for all patients admitted to the ICUs were obtained until their discharge or death. We collected only direct healthcare-related costs, defined as all costs related to the ICU stay.Enrolled patients were assessed daily in terms of cost-related expenditures such as hospital fees, operating room fees, gas therapy, physiotherapy, blood components transfusion, medications, renal replacement therapy, laboratory analysis and imaging. Standard unit costs (year 2006 values) were based on the Brazilian Medical Association (AMB) price index for medical procedures and the BRASINDICE price index for medications, solutions and hospital consumables. Medical resource utilization was also assessed daily using the Therapeutic Intervention Scoring System (TISS-28). Indirect costs were not included. RESULTS: With a mean (standard deviation [SD]) age of 61.1 +/- 19.2 years, 524 septic patients from 21 centres were included in this study. The overall hospital mortality rate was 43.8%, the mean Acute Physiology And Chronic Health Evaluation II (APACHE II) score was 22.3 +/- 5.4, and the mean Sequential Organ Failure Assessment (SOFA) score at ICU admission was 7.5 +/- 3.9.The median total cost of sepsis was $US 9632 (interquartile range [IQR] 4583-18 387; 95% CI 8657, 10 672) per patient, while the median daily ICU cost per patient was $US 934 (IQR 735-1170; 95% CI 897, 963). The median daily ICU cost per patient was significantly higher in non-survivors than in survivors, i.e. $US 1094 (IQR 888-1341; 95% CI 1058, 1157) and $US 826 (IQR 668-982; 95% CI 786, 854), respectively (p < 0.001). For patients admitted to public and private hospitals, we found a median SOFA score at ICU admission of 7.5 and 7.1, respectively (p = 0.02), and the mortality rate was 49.1% and 36.7%, respectively (p = 0.006). Patients admitted to public and private hospitals had a similar length of stay of 10 (IQR 5-19) days versus 9 (IQR 4-16) days (p = 0.091), and the median total direct costs for public ($US 9773; IQR 4643-19 221; 95% CI 8503, 10 818) versus private ($US 9490; IQR 4305-17 034; 95% CI 7610, 11 292) hospitals did not differ significantly (p = 0.37). CONCLUSIONS: The present study provides the first economic analysis of direct costs of sepsis in Brazilian ICUs and reveals that the cost of sepsis treatment is high. Despite similar ICU management, there was a significant difference regarding patient outcome between private and public hospitals. Finally, the median daily costs of non-survivor patients were higher than survivors during ICU stay.