Abstract Objective: To determine whether the greater daily expense of administering total parenteral nutrition (TPN) via plastic bags changed once daily, compared to glass bottles changed thrice daily, could be offset by savings from a reduction in nosocomial infections. Design: The costs and potential benefits of commercially available TPN bags and TPN in glass containers were compared. Costs were computed from the viewpoint of the hospital, first in a general model and then for two specific examples, Crohn's disease and intensive-care unit (ICU) patients. The extra cost of using bags was $20 per day. The total cost of nosocomial bacteremia was estimated at $6,000. The monetary benefits of using TPN bags were $6,000X T , where X T was the percentage of nosocomial infections averted. We also considered that reduction in intravenous (IV)-line manipulation could reduce bacteremia-related mortality and computed a cost-per-life-saved ratio. Results: Modeling showed that TPN in bags could yield a net benefit when the absolute reduction in the daily risk of nosocomial bacteremia reached the threshold value of 0.3%. Such a reduction could not be attained in patients with Crohn's disease, and corresponded to a 50% to 60% reduction of infection rates in ICU patients. Varying the risk of mortality attributable to IV-line-related infection from 1% to 13% resulted in a cost effectiveness of using TPN bags ranging from $90,000 to $7,000 per life saved in ICU, assuming a two-thirds reduction in IV-line infections, and from $180,000 to $14,000 if the infection rate was reduced by one third. Conclusion: The baseline cost-minimization analysis concluded that the extra cost of TPN bags was not justified by the extra savings. The cost-effectiveness analysis, however, found that the cost per life saved fell within the accepted range of public health interventions, provided a large fraction of infections are averted using TPN bags.