Implementation of early goal-directed therapy for severe sepsis and septic shock: A decision analysis

Objective: Early goal-directed therapy (EGDT) reduced mortality from septic shock in a single-center trial. However, implementation of ENT faces several barriers, including perceived costs and logistic difficulties. We conducted a decision analysis to explore the potential costs and consequences of ENT implementation. Design. Estimates of effectiveness and resource use were based on data from the original trial and published sources. Implementation costs and lifetime projections were modeled from published sources and tested in sensitivity analyses. We generated incremental cost-effectiveness ratios from the hospital (short-term) and U.S. societal (lifetime) perspectives, excluding nonhealthcare costs, and applying a 3% annual discount. Setting. Simulation of an average U.S. emergency department. Patients. Total of 1,000 simulation cohorts (n = 263 for each cohort) of adult patients with severe sepsis/septic shock. Interventions. EGDT under three alternative implementation strategies: emergency department-based, mobile intensive care unit team, and intensive care unit-based (after emergency department transfer). Measurements and Main Results. For an average emergency department, we estimated 91 cases per yr, start-up costs from $12,973 (intensive care unit-based) to $26,952 (emergency department-based), and annual outlay of $100,113. ENT reduced length of stay such that net hospital costs fell similar to 22.9% ($8,413-$8,978). EGDT implementation had a 99.4% to 99.8% probability of being dominant (saved lives and costs) from the hospital perspective, and cost from $2,749 (intensive care unit-based) to $7019 (emergency department-based) per quality-adjusted life-yr with 96.7% to 97.7% probability of being <$20,000 per quality-adjusted life-yr from the societal perspective. The intensive care unit-based strategy was the least expensive, because of lower start-up costs, but also least effective, because of implementation delay, and all three strategies had similar cost-effectiveness ratios. Sensitivity analyses showed these estimates to be particularly sensitive to EGDT's effect on mortality and intensive care unit length of stay, but insensitive to other variables. Conclusions., EGDT has important start-up costs, and modest delivery costs, but assuming LOS and mortality are reduced, EGDT can be cost-saving to the hospital and associated with favorable lifetime cost-effectiveness projections.