445. Bloodstream Infection Rates and Costs Associated with Compounded versus Multi-Chamber Bag Parenteral Nutrition in Oncology
Session: Poster Session: Hospital-acquired and Transplant Infections
Friday, October 30, 2009: 12:00 AM
Room: Poster Hall A
Background: While blood-stream infections (BSI) with parenteral nutrition (PN) are reported at 7.2-39%, they haven’t been studied in high-risk oncology populations. It is unknown whether the method of PN preparation impacts BSI rates. Our study compared PN-associated BSI rates & related hospital costs for oncology patients receiving PN sourced by compounding or multi-chamber bag (MCB).
Methods: We performed an analysis of Premier’s US database (hospital n=186); including all inpatients > 18 years, oncology ICD-9 codes 140-208, receiving PN Jan 05-Dec 07 (n = 19,540). Patients were either MCB (PN in a 2-chamber bag with glucose & amino acids; lipids & micronutrients added; n = 1,032) or COM (compounded PN; n = 18,508). Chi-square, t-tests and logistic regression examined the impact of hospital and patient characteristics (e.g., age, co-morbidities, surgery, ICU stay, APR-DRG severity, days on PN) on BSI rates (ICD-9 038.x, 790.7, 999.31 & 996.62). Linear regression estimated BSI-related hospital costs.
Results: COM patients were more likely than MCB to have major/extreme illness severity (84 v. 80%), more days PN (8.5 v. 6.1), longer ICU (4.3 v. 3.1 days) and hospital length of stay (18.3 v. 15.9 days); all p<0.001. The raw BSI rate was significantly higher for COM v. MCB (23% v. 18%; p<0.001). After accounting for baseline differences, the adjusted probability for BSI was 19% higher for COM v. MCB (23 v. 18%; odds=1.37; 95% CI=1.12 & 1.67). BSI-related costs were US$18,280 per infection. Given the above, if MCB was provided instead of COM, 894 potential BSIs may have been avoided. Assuming 80% of COM patients were appropriate for MCB, US$706 potential savings per patient may have been realized if MCB was utilized.
Conclusion: Oncology patients on COM were more severely ill, with greater LOS and a higher raw BSI rate than MCB patients. After accounting for baseline differences between COM and MCB, the adjusted probability of BSI was significantly lower for patients receiving PN via multi-chamber bag than by compounding, demonstrating increased cost-effectiveness for MCB.
Matthew Reynolds, PhD1, Victor Rosenthal, MD2, Shane Scott, PharmD3, Robin Turpin, PhD3, Katie Williams, MPH4 and  V. D. Rosenthal,
Baxter Healthcare Role(s): Collaborator, Consultant, Research Relationship, Scientific Advisor (Review Panel or Advisory Committee), Received: Research Support, Consulting Fee.
R. S. Turpin,
Baxter Healthcare Role(s): Employee, Received: Salary.
S. Scott,
Baxter Healthcare Role(s): Employee, Received: Salary.
K. Williams,
Baxter Healthcare Role(s): Grant Investigator, Research Contractor, Received: Research Support.
M. Reynolds,
Baxter Healthcare Role(s): Collaborator, Research Contractor, Research Relationship, Received: Research Support., (1)United BioSource, Lexington, MA, (2)International Nosocomial Infection Control Consortium, Buenos Aires, Argentina, (3)Baxter Healthcare, Deerfield, IL, (4)Baxter Healthcare, Lexington, MA