Identifying ICU patients who would not benefit from total parenteral nutrition

This study attempted to improve patient selection for total parenteral nutrition (TPN) by identifying ICU patients whose prognoses were so hopeless that they would not or would no longer benefit from TPN. Computerized trend analysis of daily Apache II scores, corrected for the presence, number, and duration of major organ system failures was used to predict individual hospital outcome among 50 intensive care unit patients treated with TPN. Thirteen of the 18 deaths were correctly predicted without any false predictions. During the study, the results of predictions were not used to influence clinical decision making. If TPN was withheld in the five patients predicted to die before starting TPN, or withdrawn in the eight patients when they were predicted to die, $28,350 or 11.5% of the cost of TPN during the study period would have been saved. This hypothetical study shows that elimination of inappropriate treatment of patients with hopeless prognosis with the aim of avoiding the prolongation of the process of dying and unnecessary suffering is a compassionate patient-selection strategy that can result in major savings in healthcare resources.